JP2016061653A - Headlight tester and confronting method of the same - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a headlight tester and a confronting method for performing the confronting on the basis of a lamp image imaged by receiving incident light from a headlight.SOLUTION: An image of a headlight is imaged by receiving the incident light from the headlight, and the image is binarized. By applying morphology processing to the binarized image, a shaped image separate from another image part is formed near the lamp center. The center of the shaped image is determined as a confronting center, and a headlight tester is moved so as to match the optical axis of the headlight tester with the confronting center.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a headlight tester for testing a headlight of a vehicle such as an automobile, and more particularly, a headlight tester having a function of automatically facing a headlight with respect to the headlight, and the headlight tester It is related to the facing method.

  Conventionally, most commercial headlight testers using image processing use a monochrome camera. Under such circumstances, the present applicants have previously proposed a headlight tester capable of performing color image processing on the projection light of the headlight using a color camera (Japanese Patent Laid-Open No. 2013). -2969). By using a color camera (color image sensor), it is possible to analyze the color image of the light projected from the headlight, and because the amount of information is abundant compared to using a monochrome camera, It is possible to perform various analyzes that were impossible with a conventional headlight tester using a monochrome camera.

  The headlight test is originally performed in a state where it is projected onto an external screen (also referred to as a 10m screen) located 10m ahead. In the headlight tester, an optical system and an internal system are installed in the light receiving unit. A screen is arranged and projected on the internal screen through the optical system. Since the projection light from the headlight is projected onto the internal screen via the optical system in this way, particularly when performing color image processing, it is possible to correct an error due to the influence of chromatic aberration or the like of the optical system. It becomes important.

  For example, there are halogen, HID, LED, and the like as the light source type of the headlight, but it is impossible in the prior art to automatically determine the light source type using the projection light from the headlight. Furthermore, there is no practical technique for performing color image processing on the projection light from the headlight to measure the intensity of the headlight, the optical axis (irradiation direction), and detect the elbow point. Further, when determining the cut line of the low light, it is necessary to search for the light / dark branch line of the light distribution pattern, but the search technique for the light / dark branch line in the headlight tester still requires improvement.

  Furthermore, as for the apparatus and method for directly facing the headlight tester with respect to the headlight to be tested by determining the center position of the headlight by image processing, the technique applying the elliptic approximation is first applied to the present applicant, etc. (Japanese Patent Laid-Open No. 2013-2906). However, there are generally three types of headlight shape types: projector type, multi-reflector type, and lens cut type. The prior art based on this elliptical approximation is a projector type headlight. However, it is not applicable to projector-type headlights, and it is necessary to determine the center position of the headlights separately based on circular approximation. Therefore, this prior art is not uniformly applicable to all types of headlights.

JP 2013-2969 A JP 2013-2906 A

  The present invention has been made in view of the above points, and it is an object of the present invention to provide a headlight tester capable of eliminating the above-described drawbacks of the prior art and performing a test with high accuracy and a method for facing the headlight tester. And Furthermore, another object of the present invention is to provide a headlight tester capable of performing a test quickly and with high accuracy by receiving projection light from a headlight through an optical system and performing image processing. The object is to provide a method for facing the person.

  The headlight tester of the present invention and the method of facing the headlight tester of the present invention can take various forms, but the headlight tester is directly faced to the headlight by processing the image of the headlight. The basic feature is that it is equipped with unique facing means and processes. Therefore, the basic configuration of the headlight tester is to have a front-facing imaging device for processing the headlight image. As the front-facing imaging device, either a color camera or a monochrome camera can be used, and as the imaging device of the camera, either a CCD image sensor or a CMOS image sensor can be used. is there. In one preferred embodiment described below, from the viewpoint of cost and processing speed, a monochrome camera is used as the facing imaging device and a CMOS image sensor is used as the imaging device. Should not be limited to only such specific embodiments.

  Furthermore, the headlight tester of the present invention projects the light from the headlight to determine at least one of the optical axis (irradiation direction), luminous intensity, and elbow point (light / dark branch line) of the light projected from the headlight. An optical system that receives light, a screen that projects a light distribution pattern of the projection light, a measurement imaging device that images the light distribution pattern on the screen, and receives image data from the measurement imaging device and performs image processing In some cases, an image processing apparatus that determines at least one of the optical axis, the luminous intensity, and the elbow point of the projection light may be included. As the measurement imaging device, either a color camera or a monochrome camera can be used, and as the imaging device of the camera, either a CCD image sensor or a CMOS image sensor can be used. is there. In one preferred embodiment described below, a color camera is used as the measurement imaging device and a CCD image sensor is used as the imaging device, but the present invention is limited to such a specific embodiment. Should not be limited to.

  First, various aspects when a color camera for measurement is used in the headlight tester of the present invention will be described below.

I. Determination of Light Source Type According to one aspect of the present invention, an optical system that receives projection light from a headlight, a screen on which a light distribution pattern of the projection light is projected via the optical system, and the light distribution on the screen A measurement color camera that captures a pattern, a light source type determination unit that receives RGB data from the measurement color camera and determines a light source type of the headlight based on a relationship between at least two of the RGB data; A headlight tester is provided.

  According to one preferred embodiment, the light source type is determined based on the ratio between two selected RGB data. According to yet another preferred embodiment, the light source type is determined based on the difference between the two data. According to yet another embodiment, the light source type is determined based on both the ratio and difference between the two data. According to one embodiment, the two data are G and R data, the ratio is G / R, and the difference is | G−R |.

  According to one preferred embodiment, the determination of the light source type is a determination of a warm color light source (halogen as a representative example) and a cold color light source (HID as a representative example).

II. Determination of Optical Axis According to another aspect of the present invention, an optical system that receives projection light from a headlight, a screen on which a light distribution pattern of the projection light is projected via the optical system, and the distribution on the screen. A measurement color camera for imaging a light pattern, rgb coefficient determination means for determining an rgb coefficient for each of a plurality of selected pixels of the measurement color camera, RGB data received from the measurement color camera and the rgb Luminance determination means for determining the luminance for the corresponding pixel by applying the rgb coefficient determined for the corresponding pixel determined by the coefficient determination means; discriminating the highest luminance among the luminances and determining the vertical and horizontal positions of the highest luminance There is provided a headlight tester having optical axis determining means for determining the optical axis of the headlight.

  According to one preferred embodiment, the optical system has a main lens. More preferably, the main lens is a Fresnel lens. In one preferred embodiment, the measurement color camera has a color image sensor having a predetermined number of pixels, and RGB data for each pixel of the color image sensor is output. Preferably, the color image sensor is a single-plate type, and the RGB color filter has a so-called Bayer arrangement, and the color camera for measurement performs interpolation processing to calculate respective RGB data for each pixel. Output. In another preferred embodiment, the color image sensor is a multi-plate type, and the measurement color camera outputs RGB data directly from each pixel. In one embodiment, the color image sensor is a CCD color image sensor.

  According to one preferred embodiment, the luminance is determined as the Y coordinate in the XYZ coordinate system defined by the tristimulus values X, Y, Z in the CIE • XYZ color system, ie Y = r × R + g × G + b × B. Where R, G, and B are RGB data output from the color camera for measurement, and r, g, and b are coefficients as respective weights of the RGB data.

  When determining the rgb coefficient, it is necessary to determine the color space to which it applies. In the xy chromaticity diagram, there are several color spaces (so-called color triangles) such as sRGB, AdobeRGB, and NTSC_RGB. In one embodiment, the NTSC_RGB color space is used.

  By the way, the rgb coefficient (also abbreviated as Y coefficient) varies depending on the color temperature (K: Kelvin) of the headlight as a light source. The color temperature is a numerical value representing the color of light and is different from the luminous intensity (cd). The color tone of the light source changes in order of yellow → white → blue as the color temperature increases. That is, when the color temperature changes, the spectral distribution of light also changes, and as a result, the rgb coefficient also changes. Therefore, the rgb coefficient used when determining the luminance Y needs to correspond to the color temperature of the headlight. When the color temperature is known (measurement by a color temperature measuring device, etc.), it is possible to use the rgb coefficient corresponding to the measured color temperature, but when the color temperature is unknown, the head It is necessary to estimate and approximate the color temperature of the light.

  More preferably, the rgb coefficient is determined by correcting variation in each pixel of the color image sensor due to the influence of chromatic aberration or the like via the optical system. In general, as the color image sensor moves away from the center to the outside, the variation tends to become more prominent. In one embodiment, the rgb coefficient includes at least one reference rgb coefficient determined based on the color temperature of the CIE standard light source and at least one corrected rgb coefficient modified from the reference rgb coefficient. ing. In one embodiment, in order to absorb this variation, an rgb coefficient is determined for each pixel of the color image sensor. In another embodiment, all the pixels of the color image sensor are divided into zones each having a predetermined number of pixels, and a common rgb coefficient is applied to the pixels in each zone.

  In order to determine the maximum luminance, it is possible to apply a balance method or a binarized centroid method. In the balance method, in one embodiment, the 3 degree 30 minute method is applied.

  In one embodiment, the optical axis determination means has a maximum brightness (for example, balance point) position correction table, and a maximum brightness position correction predetermined for the position of the pixel extracted as the maximum brightness. The final maximum luminance position is determined by applying the hv position correction data. Here, h is position correction data in the horizontal (left / right) direction, and v is position correction data in the vertical (up / down) direction.

  In one preferred embodiment, a light source type determination unit is further provided, and an rgb coefficient to be applied is determined according to the determined light source type.

  In one preferred embodiment, the rgb coefficient determination means has a plurality of tables for calculating rgb coefficients for each predetermined light intensity or color temperature, and selects and uses the tables according to the measured or calculated light intensity or color temperature. Determine the rgb coefficient to be used.

  Preferably, the rgb coefficient determination unit, the luminance determination unit, and the optical axis determination unit are provided in an image processing apparatus connected to the measurement color camera in the headlight tester. In one embodiment, the image processing apparatus includes at least a computer system having a CPU and a memory. The rgb coefficient determination unit, the luminance determination unit, and the optical axis determination unit are included in the headlight tester. It is stored in the memory as part of a program for controlling the operation.

III. Determination of luminous intensity (part 1)
According to still another aspect of the present invention, an optical system that receives projection light from a headlight, a screen on which a light distribution pattern of the projection light is projected through the optical system, and the light distribution pattern on the screen Measurement color camera for imaging, rgb coefficient determination means for determining an rgb coefficient for each of a plurality of selected pixels of the measurement color camera, RGB data received from the measurement color camera and the rgb coefficient determination means A luminance determining means for determining the luminance for the corresponding pixel by applying the rgb coefficient for the corresponding pixel determined by the step of determining a maximum luminance among a plurality of luminances corresponding to each of the plurality of pixels; There is provided a headlight tester having first luminous intensity determining means for determining luminous intensity of the headlight from the maximum luminance.

  According to one preferred embodiment, the optical system has a main lens. More preferably, the main lens is a Fresnel lens. In one preferred embodiment, the measurement color camera has a color image sensor having a predetermined number of pixels, and RGB data for each pixel of the color image sensor is output. Preferably, the color image sensor is a single-plate type, and the RGB color filter has a so-called Bayer arrangement, and the color camera for measurement performs interpolation processing to calculate respective RGB data for each pixel. Output. In another preferred embodiment, the color image sensor is a multi-plate type, and the measurement color camera outputs RGB data directly from each pixel. In one embodiment, the color image sensor is a CCD color image sensor.

  According to one preferred embodiment, the luminance is determined as the Y coordinate in the XYZ coordinate system defined by the tristimulus values X, Y, Z in the CIE • XYZ color system, ie Y = r × R + g × G + b × B. Where R, G, and B are RGB data output from the color camera for measurement, and r, g, and b are coefficients as respective weights of the RGB data.

  When determining the rgb coefficient, it is necessary to determine the color space to which it applies. In the xy chromaticity diagram, there are several color spaces (so-called color triangles) such as sRGB, AdobeRGB, and NTSC_RGB. In one embodiment, the NTSC_RGB color space is used.

  By the way, the rgb coefficient (also abbreviated as Y coefficient) varies depending on the color temperature (K: Kelvin) of the headlight as a light source. The color temperature is a numerical value representing the color of light and is different from the luminous intensity (cd). The color tone of the light source changes in order of yellow → white → blue as the color temperature increases. That is, when the color temperature changes, the spectral distribution of light also changes, and as a result, the rgb coefficient also changes. Therefore, the rgb coefficient used when determining the luminance Y needs to correspond to the color temperature of the headlight. When the color temperature is known (measurement by a color temperature measuring device, etc.), it is possible to use the rgb coefficient corresponding to the measured color temperature, but when the color temperature is unknown, the head It is necessary to estimate and approximate the color temperature of the light.

  More preferably, the rgb coefficient is determined by correcting variation in each pixel of the color image sensor due to the influence of chromatic aberration or the like via the optical system. In general, as the color image sensor moves away from the center to the outside, the variation tends to become more prominent. In one embodiment, the rgb coefficient includes at least one reference rgb coefficient determined based on the color temperature of the CIE standard light source and at least one corrected rgb coefficient modified from the reference rgb coefficient. ing. In one embodiment, in order to absorb this variation, an rgb coefficient is determined for each pixel of the color image sensor. In another embodiment, all the pixels of the color image sensor are divided into zones each having a predetermined number of pixels, and a common rgb coefficient is applied to the pixels in each zone.

  In one embodiment, when determining a pixel having the highest luminance, it is possible to apply a balance method or a binarized centroid method. In the balance method, in one embodiment, the 3 degree 30 minute method is applied.

  In one embodiment, the first luminous intensity determining means is predetermined between the luminous intensity of a reference light source (eg, a standard headlight or a reference sphere) having a known luminous intensity and the luminance determined by the luminance determining means. The luminous intensity is determined based on the conditions. In one embodiment, a reference light source (eg, standard headlight or reference sphere) having a known luminous intensity is used to adjust the brightness of the projected light of that reference light source (eg, standard headlight or reference sphere). Measure and preset the relationship between luminous intensity and luminance.

  In one preferred embodiment, a light source type determination unit is further provided, and an rgb coefficient to be applied is determined according to the determined light source type.

  In one preferred embodiment, the rgb coefficient determination means has a plurality of tables for calculating rgb coefficients for each predetermined light intensity or color temperature, and selects and uses the tables according to the measured or calculated light intensity or color temperature. Determine the rgb coefficient to be used.

  Preferably, the rgb coefficient determination means, the luminance determination means, and the first luminous intensity determination means are provided in an image processing apparatus connected to the measurement color camera in the headlight tester. In one embodiment, the image processing apparatus includes at least a computer system having a CPU and a memory, and the rgb coefficient determination unit, the luminance determination unit, and the first luminous intensity determination unit include the headlight tester. Is stored in the memory as a part of a program for controlling the operation of.

IV. Determination of luminous intensity (part 2)
According to still another aspect of the present invention, an optical system that receives projection light from a headlight, a screen on which a light distribution pattern of the projection light is projected through the optical system, and the light distribution pattern on the screen A measurement color camera for imaging, rgb coefficient determination means for determining an rgb coefficient for each of a plurality of selected pixels of the measurement color camera, RGB data received from the measurement camera, and the rgb coefficient determination means Luminance determining means for determining the luminance for the corresponding pixel by applying the rgb coefficient for the determined corresponding pixel, determining the luminance at a predetermined position from the center of the headlight in the light distribution pattern, and from the determined luminance A headlight tester having a second luminous intensity determining means for determining the luminous intensity of the headlight is provided.

  According to one preferred embodiment, the optical system has a main lens. More preferably, the main lens is a Fresnel lens. In one preferred embodiment, the measurement color camera has a color image sensor having a predetermined number of pixels, and RGB data for each pixel of the color image sensor is output. Preferably, the color image sensor is a single-plate type, and the RGB color filter has a so-called Bayer arrangement, and the color camera for measurement performs interpolation processing to calculate respective RGB data for each pixel. Output. In another preferred embodiment, the color image sensor is a multi-plate type, and the measurement color camera outputs RGB data directly from each pixel. In one embodiment, the color image sensor is a CCD color image sensor.

  According to one preferred embodiment, the luminance is determined as the Y coordinate in the XYZ coordinate system defined by the tristimulus values X, Y, Z in the CIE • XYZ color system, ie Y = r × R + g × G + b × B. Where R, G, and B are RGB data output from the color camera for measurement, and r, g, and b are coefficients as respective weights of the RGB data.

  When determining the rgb coefficient, it is necessary to determine the color space to which it applies. In the xy chromaticity diagram, there are several color spaces (so-called color triangles) such as sRGB, AdobeRGB, and NTSC_RGB. In one embodiment, the NTSC_RGB color space is used.

  By the way, the rgb coefficient (also abbreviated as Y coefficient) varies depending on the color temperature (K: Kelvin) of the headlight as a light source. The color temperature is a numerical value representing the color of light and is different from the luminous intensity (cd). The color tone of the light source changes in order of yellow → white → blue as the color temperature increases. That is, when the color temperature changes, the spectral distribution of light also changes, and as a result, the rgb coefficient also changes. Therefore, the rgb coefficient used when determining the luminance Y needs to correspond to the color temperature of the headlight. When the color temperature is known (measurement by a color temperature measuring device, etc.), it is possible to use the rgb coefficient corresponding to the measured color temperature, but when the color temperature is unknown, the head It is necessary to estimate and approximate the color temperature of the light.

  More preferably, the rgb coefficient is determined by correcting variation in each pixel of the color image sensor due to the influence of chromatic aberration or the like via the optical system. In general, as the color image sensor moves away from the center to the outside, the variation tends to become more prominent. In one embodiment, the rgb coefficient includes at least one reference rgb coefficient determined based on the color temperature of the CIE standard light source and at least one corrected rgb coefficient modified from the reference rgb coefficient. ing. In one embodiment, in order to absorb this variation, an rgb coefficient is determined for each pixel of the color image sensor. In another embodiment, all the pixels of the color image sensor are divided into zones each having a predetermined number of pixels, and a common rgb coefficient is applied to the pixels in each zone.

  In one embodiment, the second luminous intensity determining means determines the luminous intensity based on a predetermined condition between the luminous intensity of the headlight and the luminance determined by the luminance determining means. In one embodiment, a reference light source (eg, standard headlight or reference sphere) having a known luminous intensity is used to adjust the brightness of the projected light of that reference light source (eg, standard headlight or reference sphere). Measure and preset the relationship between luminous intensity and luminance. The predetermined position from the center of the headlight is, in one embodiment, a so-called road surface irradiation point when the headlight is in the passing mode, which is 1.3 degrees left and 0.6 degrees below the lamp center. (When the headlight height is 1 m or less) or a point at 1.3 degrees to the left and 0.9 degrees below (when the headlight height exceeds 1 m).

  In one preferred embodiment, a light source type determination unit is further provided, and an rgb coefficient to be applied is determined according to the determined light source type.

  In one preferred embodiment, the rgb coefficient determination means has a plurality of tables for calculating rgb coefficients for each predetermined light intensity or color temperature, and selects and uses the tables according to the measured or calculated light intensity or color temperature. Determine the rgb coefficient to be used.

  Preferably, the rgb coefficient determining means, the luminance determining means, and the second luminous intensity determining means are provided in an image processing apparatus connected to the measurement color camera in the headlight tester. In one embodiment, the image processing apparatus includes at least a computer system having a CPU and a memory, and the rgb coefficient determination unit, the luminance determination unit, and the second luminous intensity determination unit include the headlight tester. Is stored in the memory as a part of a program for controlling the operation of.

V. Determination of Elbow Point According to still another aspect of the present invention, an optical system that receives projection light from a headlight, a screen on which a light distribution pattern of the projection light is projected via the optical system, the screen on the screen A measurement color camera for imaging a light distribution pattern, rgb coefficient determination means for determining an rgb coefficient for each of a plurality of selected pixels of the measurement color camera, and RGB data received from the measurement color camera; luminance determining means for determining the luminance for the corresponding pixel by applying the rgb coefficient for the corresponding pixel determined by the rgb coefficient determining means, and comparing at least one horizontal cut line with the luminance for the plurality of pixels; And at least one oblique cut line having a predetermined angle with respect to the horizontal cut line A headlight tester having elbow point determining means for determining an elbow point as an intersection is provided.

  According to one preferred embodiment, the optical system has a main lens. More preferably, the main lens is a Fresnel lens. In one preferred embodiment, the measurement color camera has a color image sensor having a predetermined number of pixels, and RGB data for each pixel of the color image sensor is output. Preferably, the color image sensor is a single-plate type, and the RGB color filter has a so-called Bayer arrangement, and the color camera for measurement performs interpolation processing to calculate respective RGB data for each pixel. Output. In another preferred embodiment, the color image sensor is a multi-plate type, and the measurement color camera outputs RGB data directly from each pixel. In one embodiment, the color image sensor is a CCD color image sensor.

  According to one preferred embodiment, the luminance is determined as the Y coordinate in the XYZ coordinate system defined by the tristimulus values X, Y, Z in the CIE • XYZ color system, ie Y = r × R + g × G + b × B. Where R, G, and B are RGB data output from the color camera for measurement, and r, g, and b are coefficients as respective weights of the RGB data.

  When determining the rgb coefficient, it is necessary to determine the color space to which it applies. In the xy chromaticity diagram, there are several color spaces (so-called color triangles) such as sRGB, AdobeRGB, and NTSC_RGB. In one embodiment, the NTSC_RGB color space is used.

  By the way, the rgb coefficient (also abbreviated as Y coefficient) varies depending on the color temperature (K: Kelvin) of the headlight as a light source. The color temperature is a numerical value representing the color of light and is different from the luminous intensity (cd). The color tone of the light source changes in order of yellow → white → blue as the color temperature increases. That is, when the color temperature changes, the spectral distribution of light also changes, and as a result, the rgb coefficient also changes. Therefore, the rgb coefficient used when determining the luminance Y needs to correspond to the color temperature of the headlight. When the color temperature is known (measurement by a color temperature measuring device, etc.), it is possible to use the rgb coefficient corresponding to the measured color temperature, but when the color temperature is unknown, the head It is necessary to estimate and approximate the color temperature of the light.

  More preferably, the rgb coefficient is determined by correcting variation in each pixel of the color image sensor due to the influence of chromatic aberration or the like via the optical system. In general, as the color image sensor moves away from the center to the outside, the variation tends to become more prominent. In one embodiment, the rgb coefficient includes at least one reference rgb coefficient determined based on the color temperature of the CIE standard light source and at least one corrected rgb coefficient modified from the reference rgb coefficient. ing. In one embodiment, in order to absorb this variation, an rgb coefficient is determined for each pixel of the color image sensor. In another embodiment, all the pixels of the color image sensor are divided into zones each having a predetermined number of pixels, and a common rgb coefficient is applied to the pixels in each zone.

  In one embodiment, the elbow point determination means comprises an hv position correction table for elbow point correction for a selected pixel of the color image sensor, with respect to the horizontal and inclined cut lines or as their intersection. Hv position correction is applied to the determined elbow point position. Preferably, the hv position correction data is commonly applied to each zone of the color image sensor. Here, h is position correction data in the horizontal (left / right) direction, and v is position correction data in the vertical (up / down) direction.

  In one preferred embodiment, a light source type determination unit is further provided, and an rgb coefficient to be applied is determined according to the determined light source type.

  In one preferred embodiment, the rgb coefficient determination means has a plurality of tables for calculating rgb coefficients for each predetermined light intensity or color temperature, and selects and uses the tables according to the measured or calculated light intensity or color temperature. Determine the rgb coefficient to be used.

  Preferably, the rgb coefficient determination means, the luminance determination means, and the elbow point determination means are provided in an image processing apparatus connected to the measurement color camera in the headlight tester. In one embodiment, the image processing apparatus includes at least a computer system having a CPU and a memory, and the rgb coefficient determination means, the luminance determination means, and the elbow point determination means are included in the headlight tester. It is stored in the memory as part of a program for controlling the operation.

VI. Search for Bright and Dark Branch Lines According to still another aspect of the present invention, an optical system that receives projection light from a headlight, a screen on which a light distribution pattern of the projection light is projected via the optical system, A measurement color camera for imaging the light distribution pattern, rgb coefficient determination means for determining an rgb coefficient for each of a plurality of selected pixels of the measurement color camera, and receiving RGB data from the measurement color camera; Luminance determining means for determining the luminance for the corresponding pixel by applying the rgb coefficient for the corresponding pixel determined by the rgb coefficient determining means, and comparing the luminance for the plurality of pixels with each other, the brightness in the light distribution pattern There is provided a headlight tester having a light / dark branch line determining means for determining a light / dark branch line that is a boundary between a portion and a dark portion.

  According to one preferred embodiment, the optical system has a main lens. More preferably, the main lens is a Fresnel lens. In one preferred embodiment, the measurement color camera has a color image sensor having a predetermined number of pixels, and RGB data for each pixel of the color image sensor is output. Preferably, the color image sensor is a single-plate type, and the RGB color filter has a so-called Bayer arrangement, and the color camera for measurement performs interpolation processing to calculate respective RGB data for each pixel. Output. In another preferred embodiment, the color image sensor is a multi-plate type, and the measurement color camera outputs RGB data directly from each pixel. In one embodiment, the color image sensor is a CCD color image sensor.

  According to one preferred embodiment, the luminance is determined as the Y coordinate in the XYZ coordinate system defined by the tristimulus values X, Y, Z in the CIE • XYZ color system, ie Y = r × R + g × G + b × B. Where R, G, and B are RGB data output from the color camera for measurement, and r, g, and b are coefficients as respective weights of the RGB data.

  When determining the rgb coefficient, it is necessary to determine the color space to which it applies. In the xy chromaticity diagram, there are several color spaces (so-called color triangles) such as sRGB, AdobeRGB, and NTSC_RGB. In one embodiment, the NTSC_RGB color space is used.

  By the way, the rgb coefficient (also abbreviated as Y coefficient) varies depending on the color temperature (K: Kelvin) of the headlight as a light source. The color temperature is a numerical value representing the color of light and is different from the luminous intensity (cd). The color tone of the light source changes in order of yellow → white → blue as the color temperature increases. That is, when the color temperature changes, the spectral distribution of light also changes, and as a result, the rgb coefficient also changes. Therefore, the rgb coefficient used when determining the luminance Y needs to correspond to the color temperature of the headlight. When the color temperature is known (measurement by a color temperature measuring device, etc.), it is possible to use the rgb coefficient corresponding to the measured color temperature, but when the color temperature is unknown, the head It is necessary to estimate and approximate the color temperature of the light.

  More preferably, the rgb coefficient is determined by correcting variation in each pixel of the color image sensor due to the influence of chromatic aberration or the like via the optical system. In general, as the color image sensor moves away from the center to the outside, the variation tends to become more prominent. In one embodiment, the rgb coefficient includes at least one reference rgb coefficient determined based on the color temperature of the CIE standard light source and at least one corrected rgb coefficient modified from the reference rgb coefficient. ing. In one embodiment, in order to absorb this variation, an rgb coefficient is determined for each pixel of the color image sensor. In another embodiment, all the pixels of the color image sensor are divided into zones each having a predetermined number of pixels, and a common rgb coefficient is applied to the pixels in each zone.

  In one preferred embodiment, a light source type determination unit is further provided, and an rgb coefficient to be applied is determined according to the determined light source type.

  In one preferred embodiment, the rgb coefficient determining means has a plurality of first tables for calculating rgb coefficients for each predetermined light intensity or color temperature, and the first table suitable for the measured or calculated light intensity or color temperature. One table is selected and the rgb coefficient to be used is determined.

  Further, in one embodiment, the rgb coefficient determination means predetermines a rgb coefficient table for searching for a light / dark branch line including a plurality of special rgb coefficients to be used when searching for a light / dark branch line in a light distribution pattern. Are included according to the light intensity or color temperature. In one embodiment, the light / dark branch line search rgb coefficient is used to search for a light / dark branch line by increasing the b coefficient relative to the r coefficient and the g coefficient and paying attention to the blue component in the light distribution pattern. Do. In one embodiment, the light-dark branch line search rgb coefficient is left as one coefficient (preferably, the most stable g coefficient) and the other two coefficients (preferably r coefficient and b coefficient). ) Is set to zero, and only one of the coefficients is used to search for a light / dark branch line. In another embodiment, the rgb coefficient determining means multiplies each rgb coefficient of the first table by the same multiplier α to determine a light / dark branch line search rgb coefficient. The multiplier α is a value greater than 1 and 4 or less.

  In one preferred embodiment, the luminance determining means determines the luminance Y using the processed RGB data after gamma processing or natural logarithm processing of the RGB data, or after determining the luminance Y Y is gamma processed or natural logarithmized. In one embodiment, the light / dark branch line determining means determines the light / dark dividing line using a difference difference method.

  Preferably, the rgb coefficient determination means, the luminance determination means, and the light / dark branch line determination means are provided in an image processing apparatus connected to the measurement color camera in the headlight tester. In one embodiment, the image processing apparatus includes at least a computer system having a CPU and a memory, and the rgb coefficient determination means, the luminance determination means, and the elbow point determination means are included in the headlight tester. It is stored in the memory as part of a program for controlling the operation.

  Next, a headlight tester having the unique facing mechanism of the present invention or a facing method having a unique facing process will be described. Such a headlight tester or facing method is the measurement described above. It should be noted that the present invention is not necessarily limited to the configuration including the color camera, and can be applied to the configuration including the measurement monochrome camera.

VII. Basic alignment Basically, aligning (matching) the optical axis (center) of the main lens of the headlight tester with the lamp center (bulb center) of the headlight.

  According to one aspect of the present invention, an optical system that receives projection light from a headlight, a front-facing imaging device that receives the projection light through the optical system and captures an image of the headlight, An image processing device for receiving image data of the headlight from a counter imaging device, binarizing means for converting the image data into binarized data, and another image portion around the lamp center of the headlight An image processing apparatus comprising: morphological processing means for performing morphological processing on the binarized data to form a separated shaped image; and center determining means for determining the center of the shaped image, There is provided a headlight tester having movement control means for performing movement control for facing the headlight based on information on the center of the shaped image supplied from an image processing apparatus. .

  According to one preferred embodiment, the optical system has a main lens. More preferably, the main lens is a Fresnel lens.

  In one embodiment, the image processing apparatus is composed of a computer system having at least a CPU and a memory, and the binarization means, the morphology processing means, and the center determination means are the operations of the headlight tester. Is stored in the memory as a part of a program for controlling. In one embodiment, the morphological process is a closing process (a process in which erosion is performed the same number of times as expansion after performing dilation). Preferably, the closing process is performed near the center of the lamp. This is performed a plurality of times until a desired shaped image is obtained. Further, preferably, the closing process is performed so that the shaped image has a substantially square or rectangular shape. In one embodiment, the lamp center is the bulb center.

  In one embodiment, the movement control means, when facing the headlight based on information on the center of the shaped image, determines the position of the optical axis of the headlight tester and the center of the shaped image. When the difference from the position falls within a predetermined allowable range, the movement of the headlight tester is stopped as if it was substantially directly opposed. Since the position difference occurring in that case can be determined in the image processing apparatus, digital correction is performed on the premise of the position difference when performing subsequent image processing.

  In one embodiment, when the imaging device for direct reception receives the projection light of the headlight and picks up an image of the headlight, there is a darker area around the center of the lamp than the surroundings. It has the ability to capture images. Preferably, when the imaging device for direct facing receives the projection light of the headlight and picks up an image of the headlight, a phenomenon such as overexposure or blackout does not occur in the image. Has a dynamic range. In one embodiment, the facing image pickup apparatus has a high dynamic range (HDR) function, and takes an image of a headlight using the HDR function.

  In one embodiment, the headlight tester further includes a measurement color camera connected to the image processing apparatus, and the image processing apparatus includes the information from the measurement color camera and the correctness. A desired function is achieved by mutual use of information from the counter imaging apparatus. For example, the image processing apparatus further includes a light source type determination unit that determines a light source type (halogen, HID, LED) of a headlight based on RGB data from the color camera for measurement. The processing apparatus determines the light source type (lens cut, multi-reflector) of the headlight using the light source type information. In the case of a lens cut, the captured image of the headlight has a plurality of vertical stripes that are relatively thin and have a relatively small pitch, whereas in the case of a multi-reflector, the captured image of the headlight is relatively It has a plurality of vertical stripes that are thick and have a relatively large pitch. When the light source type is halogen, it is either a lens cut or a multi-reflector, while when the light source type is HID, it is either a projector or a multi-reflector. Therefore, the image processing apparatus can identify whether it is halogen or HID based on the light source type information, and further, whether it is a lens cut or a multi-reflector by examining the thickness and pitch of vertical stripes in the captured image. It is possible to make a determination.

  According to another aspect of the present invention, the projection light from the headlight is received, an image of the headlight is captured, the image is binarized, and a morphological process is performed on the binarized image. Form a shaped image separated from other image parts around the lamp center of the headlight, determine the center of the shaped image, and move the headlight tester to align the optical axis of the headlight tester with the center A headlight tester facing method comprising:

  In one embodiment, the captured image is a monochrome image, preferably captured using the high dynamic range (HDR) function of the facing imaging device. In one embodiment, the image includes a darker image portion around the lamp center of the headlight than its surrounding portion. Preferably, the dark image portion corresponds to the position of a lamp bulb, and the lamp center substantially coincides with the bulb center. In one embodiment, the morphological process is a closing process, and the closing process is repeated until the shaped image has a desired size or shape or the shaped image disappears. When performing the closing process until it disappears, store at least the image obtained in the previous process, and when it disappears, use the image obtained in the previous process as the final shaped image. use. The closing process is a process of performing erosion by the same number of times as expansion after performing expansion (dilation). In one embodiment, the closing process is repeated until the shaped image has a substantially square or rectangular shape. In one embodiment, the center of the shaped image is substantially coincident with the bulb center of the headlight, and the center is determined as the directly-facing center. In one embodiment, the optical axis is the center (optical axis) of the main lens of the headlight tester, and preferably the main lens is a Fresnel lens.

  In one embodiment, when the headlight tester is moved to align the optical axis of the headlight tester with the center of the shaped image, the difference between the position of the optical axis and the center position of the shaped image is predetermined. If it falls within the allowable range, the movement is stopped as being substantially aligned. This is a feature incorporated in the present invention in view of the speed of operation of the headlight tester and the avoidance of complication of the apparatus. That is, the movement of the headlight tester is usually performed through a mechanical mechanism, but there is always a clearance in the mechanical mechanism, and such clearance is made as small as possible. A very accurate mechanical mechanism is required to obtain the aligned state. Furthermore, in order to obtain a desired alignment state, it may be necessary to reciprocate the headlight tester up and down and left and right while checking the alignment state many times, which delays the operation. In order to eliminate such a problem, in this embodiment of the present invention, when the difference between the position of the optical axis and the center position of the shaped image falls within a predetermined allowable range, it is substantially matched. It is assumed that Since the position difference occurring in that case can be determined in the image processing apparatus, digital correction is performed on the premise of the position difference when performing subsequent image processing.

  According to one aspect of the present invention, it is possible to accurately and quickly determine at least one of the optical axis (irradiation direction), luminous intensity, and elbow point position of the headlight by performing color image processing on the projection light of the headlight. It is. Furthermore, it is possible to accurately and quickly determine the light / dark branch line in the light distribution pattern of the low-light headlight.

  According to the present invention, it is possible to face the headlight tester quickly and accurately with respect to the headlight.

BRIEF DESCRIPTION OF THE DRAWINGS The schematic front view which showed the headlight tester comprised based on one Example of this invention. (A) thru | or (D) is the schematic which respectively showed some examples of arrangement | positioning of the optical system in the light-receiving part which comprises some headlight testers of FIG. 1, and a camera. Schematic which showed the structure which performs the data processing in the headlight tester of FIG. Schematic which showed the relationship which reproduces the external screen (10m screen) in the distance of 10 m from a headlight on the internal screen in the light-receiving part of the headlight tester of FIG. FIG. 2 is a schematic diagram illustrating a relationship between an internal screen of the headlight tester of FIG. 1 and a color camera for measurement that captures a light distribution pattern of light projected from the headlight received on the internal screen. Schematic which showed the relationship between an external screen (10m screen), the internal screen of the headlight tester of FIG. 1, and the image sensor in the color camera for a measurement. The table | surface which showed one example of the RGB relative intensity ratio used when the projection light from a headlight is color-image-processed and it determines to the halogen and HID as the light source classification. A table of chromaticity coordinates of the white point of various standard illuminants published by the CIE described in Wikipedia. A table of chromaticity coordinates of the primary color RGB in the NTSC_RGB color space. The table | surface which showed the rgb coefficient calculated using the chromaticity coordinate of the white point of a CIE standard light source, and the chromaticity coordinate of the primary color RGB of NTSC_RGB color space. The graph which showed an example of the approximate expression of the rgb coefficient in the table | surface of FIG. (A) and (B) each a schematic view showing the position of the maximum luminance _{Y MAX} in the light distribution pattern 33 on the screen. (A) And (B) is each schematic which showed the principle which searches the position of the highest brightness | luminance by a balance system. (A) is a flowchart showing an example of a procedure for searching for the position of the highest luminance by the balance method, and (B) shows the processing status for the light distribution pattern 33 on the screen 19 at each stage of the flowchart of (A). Illustration. Schematic which showed the principle which searches the position of the highest brightness | luminance by the binarization centroid method. The graph which showed the value of the effective brightness | luminance Y obtained when changing shutter speed (SP) with respect to each setting light intensity (hcd) of a reference | standard sphere. (A) shows the shutter speed (SP) and the gradient of the corresponding brightness value in FIG. 16, and the A coefficient determined for each gradient by fitting the gradient to the approximate expression shown in (B). The table shown. For each shutter speed SP in FIG. 16, the value of the coefficient B corresponding to the coefficient A is determined by a linear expression Y = A × luminance + B in which the relationship between the luminance Y and the luminous intensity is defined using the gradient A and the coefficient B. Table. The schematic diagram which showed the orientation pattern of the passing lamp, its cut line, the elbow point, and the position of the road surface irradiation point. (A) is a schematic diagram showing a scan state of a pixel of an image sensor for applying a difference difference method to determine a cut line of a passing light, and (B) searches a horizontal cut line by a difference difference method. Schematic which shows the pixel arrangement | sequence used for FIG., (C) is the schematic which shows the pixel arrangement | sequence used in order to search a diagonal cut line by the difference method of a difference. The graph which showed the RGB data distribution in the horizontal direction (horizontal axis) and the vertical direction (vertical axis) in the vicinity of the lamp center when the luminous intensity of the traveling lamp reference sphere is set to 50, 400, and 1200 hcd, respectively. The graph which showed the RGB data distribution in the horizontal direction (horizontal axis) and the vertical direction (vertical axis) in the vicinity of the center of the lamp when the luminous intensity of the passing lamp reference sphere is set to 50, 100 and 120 hcd, respectively. (A) The top view of the system which measures the luminous intensity of the projection light from a reference | standard sphere with the illuminometer on a 10m screen, (B) is a side view of this system. The explanatory view showing the mode which measures the luminous intensity using the system of Drawing 23. Schematic which showed the aspect which equally divided | segmented the measurement range which consists of 640x480 pixels of an image sensor into the zone which consists of 20x20 pixels. Explanatory drawing of the table which showed the correspondence of a block number and an rgb coefficient. (A) And (B) is the graph of the correspondence table | surface and the approximate expression which each showed the relationship between the luminous intensity (hcd) at the time of using a reference | standard sphere as a light source, and the color temperature (K) corresponding to it. The measurement range of 640 × 480 pixels of the image sensor is equally divided into zones of 20 × 20 pixels, and an effective range is set in the measurement range, and a light source of 400 hcd is set in each zone within the effective range. Explanatory drawing at the time of assigning the appropriate block number with respect to luminous intensity. 29 is a correspondence table showing each block number of FIG. 28 and the value of the rgb coefficient corresponding to it. FIG. 24 is an explanatory diagram showing an aspect in which a longitudinal point sequence is obtained by sequentially performing longitudinal scanning at respective lateral positions using the system shown in FIG. 23. The table | surface which showed one example (500hcd) of the balance point position correction correction table | surface which showed the correspondence with the block number, the horizontal (left-right) direction position correction data h, and the vertical (up-down) direction position correction data v corresponding to it. Schematic which showed the state which assigned the block number for balance point position correction | amendment for 400 hcd driving lights with respect to each zone 41 of the image sensor 20b. 33 is a correspondence table showing the correspondence between the block numbers assigned to the respective zones 41 in FIG. 32 and the corresponding horizontal position correction amount h and vertical position correction amount v. In the system of FIG. 23, the position of the elbow point when a passing lamp reference sphere is attached as the reference sphere 51 and set to a selected position in the horizontal direction and the reference sphere 51 is angularly displaced in the vertical direction over a predetermined range is the headlight tester. The graph which showed the variation at the time of measuring by 10. FIG. 34 shows the horizontal variation and vertical variation of the measurement position of the elbow point when the selected position in the horizontal direction in the graph of FIG. 34 is set to 10 cm left (10 m screen equivalent) and angularly displaced in the vertical direction. This graph shows the approximate expression of each. (A) to (C) are correspondence tables of block numbers and horizontal corrections, and correspondences between block numbers and cut line corrections, respectively, for correcting variations in measurement positions of elbow points shown in FIGS. The table | surface which showed the table | surface and the correspondence table of a block number and height correction. FIG. 37 is a schematic diagram for explaining “horizontal correction”, “cut line correction”, and “height correction” in FIG. A graph showing the sensitivity (bright field and dark field) of the human eye. FIG. 26 is a correspondence table showing correspondences between numerical values of rgb coefficients to be used in the case of a light / dark branch line search and the corresponding block numbers (corresponding to the block numbers in the zone of FIG. 25). (A) is an image obtained by capturing a light distribution pattern of a passing light with a measurement camera when the Y coefficient is 1 (ie, r + g + b = 1), and (B) is a difference difference zero method for the light distribution pattern of (A). An image in which points obtained at intervals of 1 degree on the left and right are plotted. (A) is an image obtained by capturing a light distribution pattern of a passing lamp with a measurement camera when the Y coefficient is 4 (that is, 4 × r, 4 × g, 4 × b), and (B) is a light distribution of (A). An image obtained by plotting points obtained at intervals of 1 degree on the left and right sides by a zero difference method. It is the picked-up image of the various headlights imaged using the HDR function with the direct-facing imaging device, where (A) is a lens cut, (B) is a multi-reflector, (C) is another multi-reflector, (D ) Is a projector, and (E) is another multi-reflector. A lamp image captured by the front-facing imaging device. The binarized image obtained by binarizing the lamp image of FIG. The image after the 1st closing process with respect to the binarized image of FIG. Image after the second closing process. Image after the third closing process. Image after the fourth closing process. Image after the fifth closing process. Image after the sixth closing process. Image after the seventh closing process. Image after the eighth closing process. Image after the 9th closing process. Image after the 10th closing process. Image after the 11th closing process. Image after the 12th closing process. Image after the 13th closing process. Schematic which showed the state which the center of the headlight L and the center of the main lens 17 of the headlight tester entered into the predetermined tolerance, and the movement of the headlight tester was stopped. An image of the headlight before and after facing the headlight that is projected on the imaging surface of the head-to-head imaging device, between the center position of the headlight and the center position of the main lens of the headlight tester after facing the headlight tester Schematic which showed the state in which the difference in tolerance exists in. Schematic which showed the picked-up image of the headlight L which should be seen when it sees from the lamp | ramp center position.

  FIG. 1 shows a schematic front view of a headlight tester 10 constructed according to one embodiment of the present invention. The headlight tester 10 generally includes a light receiving unit 11 that receives projection light from a headlight of a vehicle such as an automobile to be inspected, and a base 12 that supports the light receiving unit 11 movably in the vertical and horizontal directions. The stand unit 14 stands upright on the base 12, and the display unit (monitor) 15 is disposed above the stand unit 14. A plurality of rolling elements 12 a are provided at the bottom of the base 12, and these rolling elements 12 a are laid on the floor 13 and extend to the left and right in FIG. 1 (not shown). ) It can move along the top. In the illustrated example, a guide bar 14b extending in the vertical direction is attached to the stand portion 14, and a bracket 14c is provided so as to be movable in the vertical direction along the guide bar 14b. The bracket 14c is provided so as to be manually rotatable around the guide rod 14b, and can be fixed to the guide rod 14b at a desired rotational position. Since the light receiving unit 11 is mounted on the bracket 14c, the light receiving unit 11 can be controlled to move up and down along the guide bar 14b and around the central axis of the guide bar 14b. A finder (telescope) 11a is attached to the light receiving unit 11, and the light receiving unit 11 can be rotated around the guide rod 14b so as to be aligned with the direction of the headlight to be inspected. Further, a first drive control unit that controls the vertical movement of the light receiving unit 11 and a second drive control unit that controls the horizontal movement of the base 12 are provided in the stand unit 14. The stand unit 14 is further provided with a console panel 14a including a plurality of switches. The operator inputs data via the console panel 14a to control the operation of the headlight tester 10. Is possible. The display unit 15 includes an LCD display that can display various inspection states and inspection results.

  In the illustrated example, a data processing device 21 is provided on the light receiving unit 11, and as will be described later, various circuits for image processing the projection light from the headlight to be inspected received by the light receiving unit 11. Elements are provided in the data processing device 21. In the illustrated example, the data processing device 21 is provided on the light receiving unit 11. However, the data processing device 21 is not necessarily provided on the light receiving unit 11. It is possible to provide the light tester 10 in other places, for example, in the stand unit 14. Further, when using a wireless connection, the light tester 10 can be provided at a position away from the headlight tester 10. The data processing device 21 only needs to be connected so as to be able to receive data from an imaging device (camera), which will be described later, provided in the light receiving unit 11 by wire or wireless. .

  2A to 2D schematically show some embodiments of the internal configuration of the light receiving unit 11. As is clear from the embodiment shown in FIG. 2A, the light receiving section 11 has a function as a housing, and a main lens 17 is provided on the front surface thereof. It is positioned so as to receive the projection light from the headlight L to be received. That is, when a headlight test is performed, it is necessary to set the distance between the headlight L to be inspected and the main lens 17 to a predetermined test distance D (usually 1 m). Therefore, usually, a vehicle such as an automobile equipped with the headlight L to be inspected is self-propelled and the headlight L is positioned at the test distance D in front of the main lens 17 of the headlight tester 10. On the other hand, the headlight tester 10 can be provided so as to be movable in the front-rear direction, and the headlight tester 10 can be set at a predetermined test distance D with respect to the headlight L to be inspected. Of course.

  The main lens 17 is usually a Fresnel lens, and is used to minimize the size of the light receiving unit 11, particularly the length thereof. That is, the headlight test is originally performed by visually projecting the projection light from the headlight onto an external screen (10 m screen) positioned 10 m ahead of the headlight. In the headlight tester, Is configured to reproduce the light distribution pattern on the 10 m screen on the screen 19 inside the light receiving unit 11 by using an optical system including a Fresnel lens. The reproduction of this 10 m screen will be described in more detail later. As shown in FIG. 2A, a pair of first and second half mirrors 22 and 18 are provided on the optical axis of the main lens 17 in the light receiving unit 11. Therefore, after the projection light from the headlight L passes through the main lens 17, a part thereof is transmitted by the first half mirror 22, and the remaining part is reflected upward. The transmitted light from the first half mirror 22 is transmitted through the second half mirror 18 to form a light distribution image of the projected light on the measurement screen 19 as an internal screen. As will be described later, the distance between the measurement screen 19 and the main lens 17 is determined based on the 10 m reproduction described above.

  A measurement camera 20 as a measurement imaging device is provided vertically above the second half mirror 18, and the measurement camera 20 distributes the projection light on the measurement screen 19 reflected by the second half mirror 18. Take an optical image. The measurement camera 20 can be a color camera or a monochrome camera. When the measurement camera 20 is a color camera, the light distribution image on the measurement screen 19 is captured as a color image. The measurement camera 20 is connected to a data processing device 21 provided on the light receiving unit 11, and data from the measurement camera 20 is supplied to the data processing device 21. On the other hand, a third mirror 23 is provided vertically above the first half mirror 22 to reflect the reflected light from the first half mirror and input it to a facing camera 24 as a facing imaging device. The facing camera 24 can be a color camera, but can also be a monochrome camera. The facing camera 24 is also attached to the light receiving unit 11, and is further connected to the data processing device 21, and data from the facing camera 24 is also supplied to the data processing device 21. The facing camera 24 is used to face the headlight L to the headlight tester 10, more specifically, the center 17 a of the main lens 17 of the light receiving unit 11. The direct-facing camera 24 is disposed in a positional relationship for capturing an image of the headlight L itself, and as described in detail later, the disposition condition is different from that of the measurement camera 20 that captures a light distribution image of projection light. Yes.

  FIG. 2B shows a configuration of another embodiment of the light receiving unit 11. The configuration of the embodiment of FIG. 2B is basically the same as the configuration of the embodiment of FIG. 2A, but the first half mirror 22 and the third configuration are the same as the configuration of the embodiment of FIG. This is the same as that obtained by removing the mirror 23 and the facing camera 24. Accordingly, the light receiving unit 11 in FIG. 2B processes only the light distribution image from the headlight L projected on the screen 19, and the light receiving unit 11 does not control the facing function. The facing operation needs to be performed using an element different from the light receiving unit 11.

  FIG. 2 (C) shows a configuration of still another embodiment of the light receiving unit 11. In this case, the measurement camera 20 and the second half mirror associated therewith from the configuration of the embodiment of FIG. 2 (A). 18 and the screen 19 are deleted. Accordingly, the light receiving unit 11 in FIG. 2C controls only the facing operation with respect to the headlight L by the facing camera 24, and does not capture the light distribution pattern of the projection light of the headlight L. FIG. 2 (D) shows the configuration of still another embodiment of the light receiving section 11, which is the longitudinal axis of the light receiving section between the facing camera 24 and the third mirror 23 in the configuration of the embodiment of FIG. 2 (C). The arrangement relationship in the direction is reversed. Accordingly, the angular arrangement of the third mirror is also changed to a state in which the reflected light from the first half mirror 22 can be reflected to the facing camera 24. Also in this case, the light receiving unit 11 manages only the facing operation with respect to the headlight L by the facing camera 24.

  FIG. 3 is a block diagram showing a connection relationship between the data processing device 21 and related components. The data processing device 21 includes an image processing device 30 and a control unit 31, and the image processing device 30 receives image data from the measurement camera 20 that captures a light distribution image, while the image processing device 30 receives the image data from the facing camera 24. The imaging data of the headlight L is received. For example, the image processing apparatus 30 is configured by a computer system including a CPU, a memory, and the like. The memory is a ROM that stores a program for controlling the operation of the headlight tester 10 and the like, and is being processed according to the program. A RAM for temporarily storing various data is included. The image processing apparatus 30 is connected to a display unit (monitor) 15 and causes the display unit 15 to display the results processed according to the program and / or various setting conditions. Further, the image processing apparatus 30 is connected to a control unit 31. The control unit 31 is connected to a console (control unit) 14a including various switches and the like for receiving an operator's input, and the light receiving unit 11 is connected. It is connected to a drive device that moves up and down, left and right, and a vertical and horizontal drive device that includes a limiter that regulates position control of the light receiving unit 11 in the vertical and horizontal directions, and a detector 14d.

  FIG. 4 shows a relational expression in the case where 10 m reproduction is performed in the light receiving unit 11, which is published by the traffic safety pollution research institute (current traffic safety environment research institute). In other words, the headlight test is essentially performed by irradiating projection light on an external screen (10 m screen) located 10 m in front of the headlight and visually observing the light distribution pattern on the 10 m screen. Is supposed to be done. The headlight tester adopts a 10 m reproduction configuration in order to effectively perform such a headlight test. That is, as shown in FIG. 4, an external screen (10 m screen) is disposed at a position A 10 m ahead of the headlight L, and the projection light from the headlight L is irradiated to the external screen (10 m). It is assumed that the light distribution pattern of the projection light of the headlight L can be visually confirmed on the screen). Therefore, when 10 m reproduction is performed by the headlight tester, the main lens 17 of the light receiving unit 11 is arranged 1 m ahead of the headlight L, and the headlight is incident on the main lens 17 by the focal length f of the main lens 17. The projection light of L is projected on the internal screen 19 disposed at the position B at the distance x from the main lens 17. At this time, the size a on the external screen (10 m screen) A becomes the size y on the internal screen B. Thus, by appropriately selecting the focal length of the main lens 17, it is possible to appropriately determine the distance x between the main lens 17 and the internal screen 19 at the position B. That is, the main lens 17 and the internal screen 19 can be disposed in the light receiving unit 11, and 10 m reproduction can be realized in the headlight tester.

  Referring to FIG. 5 in connection with FIG. 4, the measurement camera 20 captures the light distribution pattern of the projection light from the headlight L displayed on the internal screen 19 at the position B in the light receiving unit 11. It is. The measurement camera 20 includes a lens 20a and a semiconductor image sensor 20b, and the lens 20a is located at a distance z (work distance) from the internal screen 19. Accordingly, the light distribution pattern of the projection light on the internal screen 19 is imaged on the semiconductor image sensor 20b via the lens 20a, and thereby the light distribution image on the internal screen 19 is captured by the measurement camera 20. Is possible. A CCD or CMOS can be used as the semiconductor image sensor 20b.

  The relationship between FIGS. 4 and 5 is collectively shown in FIG. That is, as shown in FIG. 6, the large 10 m screen 25 originally scheduled in the headlight test is reproduced on the small internal screen 19 in the light receiving unit 11 by the focal length f of the main lens 17. The color light distribution image on the internal screen 19 is formed on the semiconductor image sensor 20b via the lens 20a of the measurement camera 20. Here, x is a 10 m correlation distance (distance between the main lens 17 and the internal screen 19), and z is a work distance between the internal screen 19 and the measurement camera 20 or its image sensor 20b.

  The semiconductor image sensor 20b in the measurement camera 20 is preferably composed of a CCD and has a two-dimensional array of a plurality of pixels as is well known. When the measurement camera 20 is a color camera, the image sensor 20b can be a single plate type or a multi-plate type. In the case of a single plate type, a plurality of pixels arranged in the shape of a two-dimensional matrix. Each of the color filters is provided with any one of the three primary colors R (red), G (green), and B (blue). In this case, the color filter array is generally a so-called Bayer array, but the present invention should not be limited to such a specific array. In the case of such an arrangement of color filters, each pixel only obtains charge corresponding to the light amount of the color component according to the corresponding filter as pixel data. Therefore, the measurement camera 20 performs A / D conversion on the obtained pixel data, and then performs color interpolation processing based on the data of a predetermined number of surrounding pixels, and performs the remaining two other lacks in each pixel. Generate primary color component data. When the single-plate type is used as the semiconductor image sensor 20b in this way, the measurement camera 20 performs the color interpolation processing as described above, thereby performing digital RGB data (levels of 0 to 255) for each pixel. (Digital data or analog data within the range) is generated and output. On the other hand, when the semiconductor image sensor 20b is a multi-plate type, RGB data can be obtained directly from each pixel. Therefore, the measurement camera 20 converts each RGB data to RGB data (level is 0 to 255). Output as digital data or analog data within the range.

  In one embodiment, as the image sensor 20b, a CCD image sensor having a measurement region made up of two-dimensionally arranged pixels of 640 (horizontal) × 480 (vertical) is used.

  Next, various embodiments when a color camera is used as the measurement camera 20 will be described in detail.

  The configuration for automatically determining the light source type of the headlight by performing color image processing based on the present invention will be described with reference to FIG. FIG. 7 summarizes the results of measurement of RGB relative intensity ratios for a number of halogen lamps and HID lamps obtained as a result of intensive studies by the present inventors. When this result is plotted with, for example, G / R on the horizontal axis and | GR | on the vertical axis, it is clear that the respective regions of the halogen lamp and the HID lamp are clearly distinguished. Therefore, if the RGB relative intensity ratios shown in FIG. 7 are stored in the image processing apparatus 30 in the form of a table or the like, the headlight L to be inspected is a halogen lamp (warm color lamp). Alternatively, the headlight tester 10 can automatically determine whether the lamp is an HID lamp (cold color lamp).

  When the conditions shown in FIG. 7 are stored as a table in the memory of the image processing apparatus 30, it is possible to take many forms, and at least the following three specific examples can be configured. It is.

(1) When only G / R is used as a parameter 0.0 ≦ G / R ≦ 1.19 → halogen lamp 1.20 ≦ G / R ≦ 1.79 → HID lamp (2) | When parameters are set: 1.0 ≦ | G−R | <30 → Halogen lamp 30 ≦ | G−R | ≦ 80 → HID lamp (3) When both G / R and | G−R | are used as parameters 0.0 ≦ G / R ≦ 1.19 and 1.0 ≦ | G−R | <30 → halogen lamp 1.20 ≦ G / R ≦ 1.79 and 1.0 ≦ | G−R | ≦ 80 → HID Lamp In the headlight tester 10, the operation for determining the type of the headlight L to be inspected will be described. First, the projection light from the headlight L to be inspected is projected onto the internal screen 19 of the light receiving unit 11. The light distribution pattern on the internal screen 19 is imaged on the image sensor 20b of the measurement camera 20, and RGB data is generated from the measurement camera 20 according to the light distribution pattern for each pixel of the image sensor 20b. Then, the highest luminance value Y in the light distribution pattern is extracted, and it is ensured that the highest luminance value Y falls within a predetermined range (for example, a range from 195 to 220).

  The relationship between the luminance value and RGB data for each pixel is expressed by the following equation.

Y = r * R + g * G + b * B
Note that RGB is RGB digital data (0 to 255) of each pixel, and rgb is a coefficient as a weight of each RGB (the rgb coefficients are collectively referred to as Y coefficients).

  In one embodiment, the measurement camera 20 has the following rgb coefficient value at a color temperature of 3170K as a default value.

r = 0.4111
g = 0.5461
b = 0.0428
Therefore, the luminance value Y of each pixel can be calculated using these rgb coefficient values and RGB data. In that case, ensuring that the extracted maximum luminance value Y falls within the predetermined range means that the measurement camera 20 can adjust the shutter speed, and the shutter speed is set appropriately. This is for setting to a proper value. That is, when the shutter speed is first set to the lower limit value, the RGB data from many pixels has a saturation value (255), so the shutter speed is sequentially increased so that the RGB data from each pixel is 255 or less. In this case, the process is initially rough, that is, the process is performed while skipping every predetermined number of pixels, and the process is performed more finely when the target value is approached. Then, it is ensured that the extracted maximum luminance value Y falls within the predetermined range.

  Based on the RGB data in the pixel corresponding to the maximum luminance value Y extracted by the above processing, it is possible to calculate the relative intensity ratio such as G / R and | GR− described above. In another embodiment, the relative intensity ratio is calculated based on a predetermined number of pixels (for example, 10 × 10 pixels) around the pixel corresponding to the highest luminance value Y thus extracted. Is also possible.

  Furthermore, as described above, when calculating the luminance value Y of each pixel, it is possible to calculate using the RGB data of each pixel, but in another embodiment, it is calculated in this way. Based on the luminance value Y of each pixel, it is also possible to replace the average value of each of the peripheral pixels (for example, 3 × 3 pixels) with the luminance value Y as the luminance value Y of the central pixel.

  Based on the relative intensity ratio calculated as described above, it is automatically determined whether the headlight L under inspection is a halogen lamp (warm color lamp) or an HID lamp (cold color lamp). Can be done.

  In the above-described embodiment, two parameters G / R and | GR | are used as parameters for determining the headlight light source type. Therefore, it is possible to select any two of these three RGB data and configure parameters such as ratio and / or difference and use them as parameters for headlight light source type determination. Of course, it is possible. Further, in the above-described embodiments, halogen lamps and HID lamps are employed as the light source types, but other light sources include LED lamps, and therefore the present invention can be applied to LED lamps. Of course there is. Further, since the color temperature when the headlight L is a halogen lamp is about 2500 to 4000 K, and the color temperature when the headlight L is a HID lamp is said to be about 4000 to 6000, the color of the headlight L It is also possible to determine the type of these light sources by measuring the temperature.

Next, based on another aspect of the present invention, the case where the optical axis, the luminous intensity, or the elbow point is measured by the headlight tester will be described. In any measurement, first, a light distribution pattern of the projection light from the headlight L is imaged by the measurement camera 20 which is a color imaging device for measurement of the headlight tester 10, and the measurement is performed according to the light distribution pattern. It is necessary for the camera 20 to supply RGB data for each pixel to the image processing device 30 and calculate the luminance Y. In one embodiment, the measurement camera supplies RGB data to the image processing device 30 as digital values from 0 to 255. In another embodiment, when the RGB data output from the measurement camera is an analog value, the analog RGB data is converted into 256 gradations (8 bits) from 0 (minimum luminance) to 255 in the image processing apparatus 30. Convert to digital RGB data of (maximum brightness). Then, the image processing device 30 processes the digital RGB data to calculate the luminance Y for each pixel of the color image sensor 20b of the measurement camera 20. In this case, the luminance Y corresponds to the Y coordinate in the XYZ coordinates defined by the tristimulus values X, Y, and Z in the CIE / XYZ color system.
Y = r * R + g * G + b * B (1)
Shall be defined by The coefficients r, g, and b are Y values when R, G, and B are each 100% in intensity, that is, a value of 1.

  Therefore, in order to calculate the luminance Y, the values of the coefficients r, g, and b as the weights of the RGB data must be determined. In order to determine the values of these coefficients r, g, and b, it is necessary to derive a conversion formula from RGB to XYZ.

  Here, the conversion formula from RGB to XYZ is represented by the following formula.

  This conversion matrix M is set as follows.

Here, x _R , y _R and z _R are the chromaticity coordinates of the primary color R (x _R = X _R / (X _R + Y _R + Z _R ) etc.), x _G , y _G and z _G are the primary colors G Chromaticity coordinates, x _B , y _B , and z _B are chromaticity coordinates of the primary color B, and C _R , C _G , and C _B are undetermined constants.

If the RGB values are normalized to [0, 1] and the white color is expressed by X _W , Y _W , and Z _W , the above equation (2) is as follows.

Since white Y _W = 1, X _W = x _W Y _W / y _W = x _W / y _W and Z _W = z _W Y _W / y _W = z _W / y _W , the above formula ( 4) is the following three simultaneous equations.

_{x W / y W = C R} x R + C G x G + C B x B
1 = C _B y _R + C _G y _G + C _B y _B (5)
_{z W / y W = C R} z R + C G z G + C B z B
In the three simultaneous equations of Equation (5), by specifying the RGB color space (for example, sRGB, AdobeRGB, NTSC_RGB, etc.), the chromaticity coordinates (x _R , y _R , x _G , y _G , x _B ) of the primary colors RGB. , Y _B ) is determined (note that, by definition, z _R = 1− (x _R + y _R ), z _G = 1− (x _G + y _G ), z _B = 1− (x _B + y _B )) Furthermore, since the chromaticity coordinates (x _W , y _W ) of the white point are specified by designating the light source (note that z _w = 1− (x _w + y _w ) by definition) It is possible to solve for indefinite constants C _R , C _G , and C _B.

From the above equation (3), Y is
_{Y = C R y R × R} + C G y G × G + C B y B × B (6)
Therefore, the coefficients r, g, and b in the definition formula (1) of the luminance Y described above are determined as follows.

r = C _R y _R , g = C _G y _G , b = C _B y _B (7)
As described above, the chromaticity coordinates of the primary colors RGB are specified by specifying the RGB color space (for example, sRGB, AdobeRGB, NTSC_RGB, etc.), and the chromaticity coordinates of the white point are specified by specifying the light source. , The r, g, and b coefficients in the definition formula (1) of the luminance Y are specified.

On the other hand, as shown in FIG. 8, the CIE (International Commission on Illumination) defines a large number of standard light sources as standard illuminants (A, B, C, D, E, F, etc.). A white point is defined as chromaticity coordinates (x, y). In FIG. 8, these white points are shown for a 2 ° field of view defined in 1931 and for a 10 ° field of view defined in 1964. In the following examples, the chromaticity coordinates (x ₂ , y ₂ ) of the white point of the 2 ° field of view defined in 1931 are used. Furthermore, in the following embodiments, assuming that NTSC_RGB is specified as the RGB color space, the chromaticity coordinates of the primary colors R, G, and B in the NTSC_RGB color space have values shown in FIG.

  Therefore, the x and y chromaticity coordinates of the NTSC_RGB color space shown in FIG. 9 and the x and y chromaticity coordinates of the white points of the respective standard light sources A to F12 shown in FIG. 8 are used. The r, g, and b coefficients for each standard light source are calculated as shown in the table of FIG. In the table of FIG. 10, the color temperature (K) for each standard light source is also described. This is a copy from the CCT (K) in the table of FIG. 8, that is, the correlated color temperature. In the present specification, the correlated color temperature is simply described as a color temperature in a superordinate concept.

  Since the table in FIG. 10 indicates that the r, g, and b coefficients change as a function of the color temperature (K), it is possible to fit an approximate expression and use a quadratic expression. An example of this is shown in FIG. Such an approximate expression is not limited to the quadratic expression shown in FIG. 11, but may be a higher order polynomial or linear expression, and may be fitted with other appropriate types of approximate expressions. Of course, it is possible. In one embodiment, the color temperature is changed using such an approximate expression to determine the r, g, and b coefficient values, and the relationship between the color temperature and the rgb coefficient is stored in the image processing apparatus 30 as a table. Can be stored. From the above results, if the color temperature of the light source is known, it is possible to specify the rgb coefficient corresponding to the color temperature, and based on the digital RGB data obtained by image processing the projection light from the light source. The luminance Y for each pixel can be determined according to the above equation (1).

  Accordingly, in the headlight tester 10 shown in FIG. 1, an apparatus capable of measuring the color temperature of the projection light from the headlight L as the light source to be measured is provided, and the measured color temperature is used as the image processing apparatus. If it is possible to supply directly to 30, it is possible to directly determine the values of the r, g and b coefficients based on the color temperature thus measured. Alternatively, as another embodiment, the color temperature of the projection light from the headlight L to be measured by the operator is measured using a device capable of measuring a separate color temperature, and the color temperature measurement value is measured. It is also possible to adopt a configuration in which an operator inputs via the operation unit 14a. In this case, an additional effort is required for the operator to separately measure and input the color temperature. However, without adopting a configuration in which the color temperature measuring device is mounted on the headlight tester 10, There is an advantage that it can be dealt with using a color temperature measuring device. On the other hand, when the color temperature measured by the color temperature measuring instrument is not used, first, the color temperature is estimated and processed, and then the estimated color temperature is to be within the required allowable range. It is necessary to take measures.

  As described above, since the luminance Y is calculated for each pixel of the color image sensor 20b of the measurement camera 20, by finding the highest luminance among the calculated luminance Y, the light of the headlight L The axis (irradiation direction) can be determined, and then the luminous intensity (hcd) corresponding to its highest brightness can be determined. On the other hand, it is also possible to determine the light intensity at a so-called road surface irradiation point at a predetermined position in the light distribution pattern by using the calculated luminance Y. It is also possible to search the horizontal cut line and the oblique cut line in the light distribution pattern by comparing these calculated luminances Y with each other and determine the position of the elbow point as the intersection of those cut lines. is there. Each of these processes will be described in detail below.

1. Determination of the optical axis (irradiation direction) of the traveling lamp As described above, if the color temperature of the headlight L as the light source is known, the rgb coefficient corresponding to the color temperature can be determined by the rgb coefficient determination means. Can be calculated by the luminance determining means. As a result, the calculated luminance Y for each pixel is temporarily stored in the RAM of the image processing apparatus 30, for example. These rgb coefficient determining means and luminance determining means are stored in, for example, a ROM as part of an image processing program in the image processing device 30 of the headlight tester 10. The image processing program further includes an optical axis determining unit as a part of the image processing program, and the optical axis determining unit has the highest luminance Y among the pixels Y temporarily stored in the RAM ( (Maximum) brightness is found, and the vertical and horizontal positions of the maximum brightness are determined as the optical axis (irradiation direction) of the projection light from the headlight L.

When the optical axis determination means finds the maximum luminance Y _MAX , a plurality of luminance Y temporarily stored in the RAM is scanned in a predetermined direction, for example, between a pair of adjacent pixels. It is possible to carry out the processing by sequentially processing so that a higher luminance Y survives, and it is also possible to apply any known algorithm for finding other maximum values. FIG. 12A schematically shows the light distribution pattern 33 projected on the measurement screen 19 in the light receiving unit 11 by a plurality of isoluminous ellipses, and the measurement camera 20 images the light distribution pattern 33. RGB data is output, and the image processing apparatus 30 processes these RGB data according to the image processing program. In the light distribution pattern 33 of FIG. 12A, the maximum luminance Y _MAX exists at the approximate center, and the position where the maximum luminance Y _MAX exists can be determined as the optical axis (irradiation direction). On the other hand, FIG. 12B shows that there are a plurality of maximum luminances Y _MAX near the center of the light distribution pattern 33, or there is only one maximum luminance Y _MAX , but this constantly changes the position near the center. Shows the case. In the case of FIG. 12B, for example, a position obtained by averaging the positions of the plurality of maximum luminous intensity Y _MAX can be determined as the optical axis (irradiation direction).

FIG. 13 shows the principle of the balance method, which is another approach for determining the position of the maximum luminance Y _MAX of the light distribution pattern 33 of the traveling light. The embodiment of FIG. 13 is a case where the 3 degree 30 minute method is applied as the balance method. By the way, as shown in the light distribution pattern 33 of FIG. 12, the luminance distribution of the light distribution pattern 33 of the traveling lamp changes almost gently both vertically and horizontally. Therefore, according to the balance method, as shown in FIG. 13A, the center position of the respective brightness of the left and right three times at a time distant two points is the same Y _H and the left and right of the center O _H, and the vertical The center of each position where the luminances of two points that are 30 minutes apart at the same Y _V are the same as the vertical center O _V. Then, the luminance at the center position determined by the center O _H of the left and right and above and below the center O _V and the maximum brightness Y _MAX.

Next, with reference to FIG. 14, a specific procedure in the case of applying the 3 degree 30 minute method as the balance method will be described in detail. FIG. 14A shows a flow chart 35 of the balance method, and FIG. 14B shows a light distribution pattern 33 on the measurement screen 19 in steps corresponding to the flow chart 35. That is, in the flowchart 35 of FIG. 14A, when the balance point measurement procedure is started, the position of the maximum luminance Y _MAX of the light distribution pattern 33 is extracted and the position is extracted in step 35a of FIG. 14A and step 36a of FIG. 14B. Set as attention position. Next, in Step 35b of FIG. 14A and Steps 36b1-36b2 of FIG. 14B, the respective luminances Y at the three left and right positions from the current target position are compared (Step 36b1), and the luminance Y at these left and right positions is balanced. The position to be updated is updated as the current attention position on the left and right (step 36b2). Next, in steps 35c of FIG. 14A and steps 36c1-36c2 of FIG. 14B, the respective luminances Y at the positions 30 minutes above and below the updated current attention position are compared (step 36c1). The position where the brightness Y balances is updated as the new upper and lower attention positions (step 36c2). Next, in step 35d, 1 is added to the count that is the number of repetitions, and then in step 35e, it is determined whether or not the count exceeds a predetermined maximum number of repetitions (for example, 50). In this case, the process returns to step 35b again to continue the balance process. On the other hand, if the count exceeds the predetermined maximum number of repetitions, the process proceeds to step 35f, where the updated current attention position in the horizontal direction and the vertical direction is used as the optical axis (irradiation direction) of the projection light of the headlight L. Determine as. In another embodiment, the iterative process is terminated when the difference between the updated target position and the target position before the update falls within a predetermined allowable range.

FIG. 15 shows a binarized centroid method, which is still another method for extracting the maximum luminance Y _MAX . According to the binarized centroid method, as shown in FIG. 15, there is a certain luminance Y (0 to 255) for each pixel obtained for the light distribution pattern 33 on the measurement screen 19. set the threshold luminance Y _T performs binarization processing to cut only the luminance of the above Y _T in. In this case, as an example, as shown in FIG. 15, Y _T is set to Y _T (= 127), which is 50% of Y _MAX (= 255), and only the luminance higher than that is cut out. . Incidentally, in the case of performing binarization at 50%, the left and right pixel positions 33d1 for obtaining the balance at 3 degrees each in the left and right direction in the balance system in which the boundary 33a of 50% applies the 3 degrees and 30 minutes system. , 33d2. Next, a cross-sectional primary moment is calculated for the cross-sectional figure of the cut boundary 33a, and the position of the center of gravity 33b 'as the center of the figure is determined. In the example of FIG. 15, it is assumed that the cross section of the luminance level in the vertical direction is symmetric with respect to the horizontal axis H, so that the position of the center of gravity 33b ′ exists on the horizontal axis H. If the cross section of the luminance level in the vertical direction is not symmetrical with respect to the horizontal axis H, the position of the center of gravity 33b 'in the vertical direction is determined by calculating the cross-sectional primary moment with respect to the cross-sectional figure of the boundary 33a in the vertical direction V. It is necessary to. In the illustrated example of FIG. 15, the determined position 33b of the center of gravity 33b ′ is slightly shifted to the left as compared to the position 33e of the true maximum luminance Y _MAX. Compared with the position 33c determined as the center of the light intensity, it is closer to the position 33e of the true maximum luminous intensity Y _MAX . This is because when the light distribution pattern 33 as shown in FIG. 15 is asymmetric in the left-right direction, the binarized centroid method is closer to the true maximum luminous intensity Y _MAX position than the balance method. The position is obtained.

2. Next, a procedure for determining the light intensity of the headlight L from the maximum luminance Y _MAX determined as described above will be described. In general, a candela (cd) is used as a unit of luminous intensity of the headlight L. However, in the headlight tester industry, a hector candela (hcd) obtained by multiplying it by 100 (ie, 1hcd = Since 100 cd) is usually used as the unit of luminous intensity, hcd or HCD is used as the unit of luminous intensity of the headlights in this specification unless otherwise specified.

  By the way, the luminance Y is merely a dimensionless numerical value obtained by performing color image processing on the RGB data obtained from the projection light of the headlight L, and is usually between luminance and luminous intensity. Since no special relationship exists, when determining the luminous intensity (hcd) from the luminance Y, the luminance and luminous intensity must be related in some way. In one embodiment of the present invention, a relationship between luminous intensity and luminance is established in advance using a reference light source (eg, standard headlight or reference sphere) whose luminous intensity is known. This will be described in detail below. In the industry, there is a reference sphere that is used to calibrate or test a headlight tester. A reference sphere is a lamp manufactured by the Japan Automobile Machine Tool Association requesting a lamp manufacturer for a standard conformity test, and in particular, a high precision manufactured to test the headlight tester itself. Lamp. The reference sphere is a glass lamp, the light distribution is a lens cut, the light source is a halogen lamp, and there are two types, one for driving lights and one for passing lamps. The reference sphere can generate a luminous intensity in the range of 50 to 12,000 hcd to be inspected by the rules.

  FIG. 16 shows a case where such a reference sphere is used to receive projection light from the reference sphere by the color image sensor 20b of the measurement camera 20, and send RGB data from the measurement camera 20 to the image processing device 30. A part of the result of outputting the luminance Y from the image processing device 30 is shown. As shown in FIG. 16, the reference sphere used has its luminous intensity (hcd) varied from 10 to 1250, initially every 5 hcd and then every 10 hcd. The measurement camera 20 can change its shutter speed (SP), and here, it is changed between SP = 0 and 64. As shown in FIG. 16, as the luminous intensity of the reference sphere increases, when the shutter speed SP is slow, the luminance Y is 255, indicating that the luminance is equal to or higher than the maximum luminance (that is, an overflow state). As the shutter speed SP is increased, an effective luminance value starts to be obtained. In this example, in order to avoid calculating a luminance value more than necessary by setting a certain threshold, the luminance value below the threshold is cut to zero. In FIG. 16, the numerical value enclosed by a square is the effective luminance Y obtained at each shutter speed SP with respect to the set luminous intensity (hcd) of the reference sphere. The table shown in the lower part of FIG. 16 is an enlarged view of a part of the upper part.

  As is apparent from the lower table of FIG. 16, the luminance Y obtained for each shutter speed SP value increases as the luminous intensity hcd of the reference sphere increases. Accordingly, the value of the luminance Y shown in FIG. 16 represents the gradient for each shutter speed SP, and the gradient defined by the value of these luminance Y for each shutter speed SP can be calculated. It is. FIG. 17A is a table showing slope values calculated for the range of shutter speeds SP = 0 to SP = 34 shown in the enlarged display portion shown in the lower part of FIG. . FIG. 17B shows an approximate expression obtained by regarding these gradient values as a function of the shutter speed SP. The approximate expression shown in FIG. 17B is an exponential function as an example, but it should be understood that any other function of any appropriate type can be applied. A value calculated for each shutter speed SP from the approximate expression thus obtained is shown in the table of FIG. The coefficient A can be said to be a smoothed gradient.

  Next, FIG. 18 is a copy of a part of the enlarged display table in the lower part of FIG. 16 (shutter speed SP is 0 to 10 and the light intensity hcd of the reference sphere is 10 to 60). The value of the coefficient A obtained with respect to the shutter speed SP shown in FIG. By the way, in the table of FIG. 18, the coefficient A represents the gradient with respect to each shutter speed SP, and for each shutter speed SP, the luminance value defining the gradient is shown over the corresponding luminous intensity hcd. Yes. Accordingly, when the relationship between the luminance and the luminous intensity at each shutter speed SP is approximated by a linear expression using the corresponding gradient (coefficient A), the following expression is obtained.

Y = A × luminance + B (8)
Therefore, for each shutter speed SP, for example, the effective maximum luminance value (for example, “242” when SP = 1) shown in the table of FIG. 18 is used as an example. It is possible to calculate the coefficient B of the linear approximation. The values of the B coefficient thus calculated are shown in the table of FIG.

  From the above, if the shutter speed SP of the measurement camera 20 when the luminance Y is obtained is known, the corresponding A coefficient and B coefficient can be obtained from the table of FIG. 18, so that the shutter speed SP is used. For the luminance Y obtained in this way, the corresponding luminous intensity hcd can be calculated by the following equation.

Luminous intensity = (Y−B) / A (9)
As described above, it is possible to determine the light intensity (hcd) as the traveling light of the headlight L corresponding to the maximum luminance Y _MAX of the traveling light. It should be noted that the manner of establishing the relationship between the luminance and the luminous intensity is not limited to the above-described embodiment, and it is needless to say that the association can be appropriately performed by other methods.

3. As described above, when the headlight L is a running light, the luminance Y is calculated for each of the selected pixels using RGB data obtained from the measurement camera 20, and then, explore the maximum brightness _{Y MAX} among those luminance Y, it was necessary to convert the maximum brightness _{Y MAX} to luminosity (hcd). On the other hand, in the case of a passing lamp, the position at which the luminous intensity is to be measured is determined in advance as a “road surface irradiation point” by a rule (the road transport vehicle law or the like). That is, as shown in FIG. 19, the road surface irradiation point 36 at which the light intensity of the passing light is to be measured is 1.3 degrees to the left of the perpendicular V passing through the center of the headlight L and below the horizontal line H passing through the center. The point 36a is 0.6 degrees (the height of the headlight L is 1 m or less) or the point 36b is 0.9 degrees (the height of the headlight L is more than 1 m). Note that a vertical line V and a horizontal line H in FIG. 19 indicate a vertical line and a horizontal line 10 m ahead of the center of each of the left and right headlights L on the extension of the center axis of the vehicle. These angles represent angles when projected 10 m ahead of the headlight L. In the light distribution pattern 33 of the passing lamp shown in FIG. 19, a cut line 37 (horizontal cut line 37a and oblique cut line 37b) and an elbow point 38 are present.

  In this way, in the measurement of the light intensity of the passing light, the position to be measured is defined invariably, so the light distribution pattern 33 in FIG. 19 is imaged by the measurement camera 20 and the RGB data is image processing apparatus 30. , The luminous intensity (hcd) of the headlight L can be determined from the luminance Y of the pixel corresponding to the position of the road surface irradiation point 36 using the above-described equation (9).

4). Detection of Elbow Point of Passing Light As described above, the luminance Y is calculated for each pixel of the color image sensor 20b of the measurement camera 20, so that these calculated luminances are shown in FIG. By comparing Y with each other, the horizontal cut line 37a and the oblique cut line 37b in the light distribution pattern 33 can be searched, and the position of the elbow point 38 can be determined as an intersection of these cut lines. In one embodiment of the present invention, a cut line is extracted using a difference difference method, and the principle of the difference difference method will be described below with reference to FIGS.

  FIG. 20A shows a light distribution pattern 33 of the projection light from the headlight L as a passing light imaged on the CCD color image sensor 20b. A light distribution pattern 33 schematically shown by a plurality of isoluminous lines is a bright portion, and a horizontal cut line 37 a is shown on the upper right side of the light distribution pattern 33, and the upper left side of the light distribution pattern 33. Shows an oblique cut line 37b. The intersection of the horizontal cut line 37a and the oblique cut line 37b constitutes an elbow point 38. The cut line 37 is a light-dark branch line that forms a boundary between the bright portion of the light distribution pattern 33 below and the dark portion above it. 20A shows a cross section of the luminance (luminance) distribution of the light distribution pattern 33 including the cut line 37. The cut line 37 has a steep curve from the bright light distribution pattern 33 to the dark portion. It exists in the maximum brightness | luminance gradient point of a brightness | luminance gradient line within a gradient part. The difference difference method is, in principle, a technique for finding the maximum luminance gradient point of the luminance gradient line by image processing.

That is, FIG. 20B shows _three measurement points e _{1 to} e 3 that are spaced apart from each other by a predetermined distance (0.44 ° pitch in the illustrated example) and arranged vertically. Each measurement point corresponds to each pixel of the image sensor 20b. If the luminance values at the respective measurement points e ₁ , e ₂ , e ₃ are Y ₁ , Y ₂ , Y ₃ , respectively,
(Y ₁ + Y ₃ ) -2Y ₂ = 0 (10)
To determine the position of the measurement point e ₂ that satisfy as the position of the cut line 37. This difference formula can be rewritten as Y ₁ −Y ₂ = Y ₂ −Y ₃ , which is the maximum gradient constituting the inflection point in the luminance gradient line in FIG. It means that it is a point. When the measurement points are arranged at a pitch of 0.44 ° as in the illustrated example, the pixel 40 of the image sensor 20b is sequentially moved downward about every 0.1 ° to perform a vertical scan 39a and a horizontal cut line. It is possible to find the position of 37a. When the scan for one pixel column is completed, the vertical scan 39a is performed in the same manner for another pixel column shifted by a predetermined distance in the horizontal direction to determine the position of the horizontal cut line 37a in the pixel column, and so on. Thus, it is possible to determine the position of the horizontal cut line 37a by performing all scanning over a preset range. That is, the horizontal cut line 37a as a point sequence by connecting multiple points locates the point of their measurement point e ₂ that satisfy the equation of the difference method of the difference by repeating the vertical scan 39a.

On the other hand, in order to search for the oblique cut line 37b, as shown in FIG. 20C, these three measurement points e _{1 to} e ₃ are shifted from each other in the horizontal direction. In the case of the illustrated example, the pixel is arranged to be shifted to the left by one pixel at the lower side by four pixels. This is basically because the oblique cut line 37b is inclined 15 degrees with respect to the horizontal cut line 37a. This arrangement takes into account In this case, the scan direction is an oblique scan 39b, and the scan is performed in a direction substantially orthogonal to the oblique cut line 37b. Even in the case of the oblique scan 39b, the difference difference method formula is applied between the _three measurement points e _{1 to} e ₃ so that the position of the pixel 40 that satisfies the difference difference method equation is scanned obliquely. It is determined as the position of the oblique cut line 37b at 39b.

  As described above, when the horizontal cut line 37a and the oblique cut line 37b are determined, it is possible to find an intersection of these cut lines, and such an intersection can be determined as the elbow point 38. It is.

  In the above-described embodiment, the pitch between the three measurement points is set to 0.44 °. However, this is merely an example. For example, the pitch is 0.23 ° (corresponding to UTAC) or the color image sensor 20b to be used. Of course, any other pitch can be selected depending on the pixel arrangement. As described above, the horizontal cut lines 37a and the oblique cut lines 37b in the projection light from the headlight L are found by performing image processing for comparing the luminances Y of the plurality of pixels 40 of the color image sensor 20b with each other. It is possible to determine the elbow point 38 as the intersection of the cut lines.

  As described above, the measurement camera 20 having the CCD color image sensor 20b is used to process the projection light from the headlight L to output RGB data, and these RGB data are converted into the color of the light source in the image processing apparatus 30. By performing image processing based on temperature, it is possible to determine the optical axis, luminous intensity, elbow point, and the like. However, the projection light of the headlight L is not irradiated on the 10 m screen ahead of 10 m, but is irradiated on the internal measurement screen 19 via the optical system including the lens 17 in the headlight tester 10. Therefore, depending on the design conditions of the headlight tester 10, there is a case where the optical system is particularly affected by chromatic aberration. The present invention is characterized in that it also includes a technique for removing the adverse effect of the optical system caused by such chromatic aberration.

  By the way, using the above-described reference sphere as a light source, the luminous intensity is set in steps of 50 hcd, 400 hcd, and 1200 hcd sequentially, and the irradiation direction is set to 0-0 (left-right direction angle 0-up-down direction angle 0). FIG. 21 (running light) shows a bitmap data obtained by imaging the light distribution pattern with the measurement camera 20 of the headlight tester 10, and collecting and plotting RGB data values from the bitmap data. FIG. 22 (in the case of a reference sphere) and FIG. 22 (in the case of a passing light reference sphere). In these drawings, the horizontal axis position is a horizontal position near the center of the light distribution pattern, and the vertical axis position is a vertical position near the center of the light distribution pattern. Furthermore, in these figures, the level is a data value for each of RGB and takes a value in the range of 0 to 255. It should be noted that, in these figures, the RGB data values (levels) are almost the same even though the luminous intensity of the reference sphere is increased from 50 hcd to 1200 hcd. However, this is because the shutter speed of the measurement camera is increased as the luminous intensity is increased due to display limitations. Therefore, absolute evaluation of RGB data values themselves is meaningless.

  In the case of the traveling light reference sphere of FIG. 21, the light intensity is relatively low at 50 hcd, the light distribution is also reddish, and the data values of G and B are even lower than the data value of R. At 400 hcd, the RGB data values are almost the same, indicating an average light distribution as a headlight. At 1200 hcd, the light intensity is high and the light distribution is whitish, and the G and B data values are reversed and larger than the R data value. In the case of the passing light reference sphere in FIG. 22, as the luminous intensity increases in order from 50 hcd to 100 hcd and 120 hcd, there is a tendency that the sizes of the respective RGB data values are reversed, although relatively little. In the case of the passing light reference sphere, since the range of changing the luminous intensity is narrower, there is no significant difference between the RGB data values. The reversal phenomenon of the respective sizes of the RGB data is more prominent in the peripheral portion than in the central portion of the image sensor 20b, which is considered to be caused by measuring the projection light through the optical system. Therefore, when the influence of the optical system to be used is large, it may be appropriate to perform some correction particularly in the peripheral portion.

5. As described above, according to one aspect of the present invention, each value of the rgb coefficient can be determined based on the color temperature of the headlight L as the light source. ) To determine the luminance Y and the corresponding luminous intensity. However, in the headlight tester 10, in order to realize 10 m reproduction, the projection light from the headlight L is imaged by the measurement camera 20 which is a color camera via an optical system including the main lens 17. Therefore, the projection light is affected by chromatic aberration or the like by passing through the optical system, and the influence may not be negligible depending on the configuration of the optical system. In particular, when a Fresnel lens is used as the main lens 17, there is a possibility that the influence of such chromatic aberration or the like becomes more remarkable.

  In order to cope with such a situation, according to one aspect of the present invention, the reference rgb coefficient determined (selected) based on the color temperature of the headlight L (ie, determined based on the CIE standard light source). rgb coefficient) is not commonly applied to all pixels of the image sensor, but at least one r, g, or b coefficient modified rgb coefficient is applied to at least one pixel. doing. For example, as described above, fluctuations in the data of RGB are relatively small in the central portion of the image sensor, but are more noticeable in the peripheral portion. Therefore, in one embodiment of the present invention, in the central portion of the image sensor, A reference rgb coefficient determined (selected) based on the color temperature of the headlight L is applied to each pixel, and a corrected rgb coefficient obtained by appropriately correcting the reference rgb coefficient is applied to each peripheral pixel. Apply. However, particularly when a Fresnel lens is used, the influence of the optical system is particularly great, and the influence of chromatic aberration or the like may not be negligible even in the central portion. In another embodiment of the present invention, the image sensor Regardless of position, the correction rgb coefficient is applied to at least one pixel or most pixels. Furthermore, in another embodiment of the present invention, instead of applying the rgb coefficient for each individual pixel, all the pixels within the measurement range of the image sensor are divided into zones composed of a plurality of pixels, and the rgb coefficient is obtained. Is commonly applied to a plurality of pixels in each zone. Therefore, for example, the reference rgb coefficient is applied to one or more zones in the central portion of the image sensor, and the corrected rgb coefficient is applied to the other zones. In this case, it is possible to prepare a plurality of correction rgb coefficients different from each other, and apply one correction rgb coefficient in common to a plurality of pixels in one zone. In this case, it is desirable to store the reference rgb coefficient and the plurality of corrected rgb coefficients in the image processing apparatus 30 in the form of a table. Further, since the reference rgb coefficient changes depending on the color temperature of the headlight L as the light source, a table of the reference rgb coefficient and a plurality of rgb correction coefficients is provided for each preselected color temperature or for each corresponding light intensity. It is desirable to prepare it and store it in the image processing apparatus 30.

  Next, a method for creating the rgb correction coefficient described above will be described.

  FIGS. 23A and 23B show a system for measuring the luminous intensity (hcd) of the projection light from the reference sphere 51 described above. The system basically includes the reference sphere mounting device 50 and 10 m therefrom. And a 10 m screen 25 that is disposed in front and includes a luminometer 58. The reference sphere mounting device 50 includes a mounting bracket 52 a to which the reference sphere 51 can be detachably mounted. The mounting bracket 52 a is fixed to the turntable table 52. The turntable table 52 is rotatably supported by a turntable 54, and the turntable table 54 is provided on a turntable front / rear moving base 56. Further, a left / right rotation handle and angle scale 53 are attached to the turntable table 52, and a vertical rotation handle and angle scale 55 are attached to the turntable 54. Although the specific configuration is not shown in FIG. 23, the turntable table 52 can be moved up and down to be fixed at a desired height.

  The reference sphere mounting device 50 is moved back and forth by the turntable forward / backward movement base 56, and the position of the rotation center 51 a of the reference sphere 51 is set to a position 10 m from the 10 m screen 25 and is fixed to that position. The 10 m screen 25 is provided with an illuminance meter 58 on the same plane. The illuminance meter 58 is special and has the same structure and function as the illuminance meter 1 shown in FIG. 1 of the aforementioned Japanese Patent Application Laid-Open No. 2013-2969. In practice, however, in order to measure the light intensity of the headlight, a candela (or hector candela) value converted as the light intensity (hcd) of the headlight L is output as the measured value. With such a system configuration, it is possible to measure the luminous intensity (hcd) by receiving the projection light from the reference sphere 51 with the illuminometer 58. In this case, the illuminometer 58 is fixed to the 10 m screen 25, but the projection direction of the reference sphere 51 can be changed in the left-right direction by operating the left-right rotation handle and the angle scale 53. Further, the projection direction of the reference sphere 51 can be changed in the vertical direction by operating the vertical rotation handle and the angle scale 55. Since the reference sphere mounting device 50 is manufactured with extremely high accuracy, changes in the vertical and horizontal directions of the reference sphere 51 can be performed with extremely high accuracy.

  Next, an aspect in which the light intensity of the reference sphere 51 is measured by the system shown in FIG. 23 will be described with reference to FIG. The positional relationship between a selected position of the 10 m screen 25 and the corresponding position of the screen 19 in the light receiving unit 11 of the headlight tester 10 and the corresponding pixel of the image sensor 20b is unique as shown in FIG. Can be determined. Therefore, for example, as shown in FIG. 24, when the projection direction of the reference sphere 51 is directed to a specific position of the lower 20 on the 10 m screen 25 and the left 35 cm / 10 m, an image corresponding to that position is displayed. It is possible to specify the pixel of the sensor 20b. That is, basically, the position of each pixel of the image sensor 20 b can be identified on the 10 m screen 25. Therefore, it is possible to determine the luminous intensity of the reference sphere 51 for each pixel of the image sensor 20b using the reference rgb coefficient corresponding to the color temperature of the reference sphere 51, while using the illuminometer 58. It is possible to determine the luminous intensity at a position on the 10 m screen 25 with respect to the position of each pixel of the image sensor 20b. The light intensity values determined via these image sensors 20b and the light intensity values determined by the illuminometer 58 are compared, and if they substantially match, the reference rgb coefficient is set for those pixels. If they do not substantially match, the light intensity values substantially matching the light intensity values obtained with the luminometer 58 are obtained for the corresponding pixels so that the image processing device 30 can obtain the light intensity values. The reference rgb coefficient is corrected to newly determine a corrected rgb coefficient.

  When determining such a correction rgb coefficient, in one embodiment, the correction rgb coefficient is determined finely for each individual pixel, and in another embodiment, all pixels within the measurement range of the image sensor 20b are determined in advance. Are divided into a plurality of zones, and a corrected rgb coefficient is determined for each zone. In the zone division embodiment, the reference rgb coefficient or the corrected rgb coefficient determined for the zone is commonly applied to a plurality of pixels belonging to one zone. When determining the corrected rgb coefficient, the value of the reference rgb coefficient is adjusted so that the difference between the light intensity obtained using the reference rgb coefficient and the light intensity obtained via the illuminometer 51 is zero. However, there are several modes for the adjustment method. That is, in one embodiment, all values of the reference rgb coefficient are adjusted to obtain a corrected rgb coefficient. In another embodiment, at least one of the reference rgb coefficients is set to zero and the rest is set. Is adjusted as appropriate to obtain a corrected rgb coefficient. When adjusting the value of the reference rgb coefficient, in one embodiment, the respective coefficients of all the reference rgb coefficients are adjusted independently of each other, and in another embodiment, the reference rgb coefficient is adjusted. Each value is commonly adjusted by multiplying the same value. Further, since the reference rgb coefficient changes depending on the color temperature of the headlight L, for each predetermined light temperature or corresponding light intensity (for example, 50, 75, 100, 150, 400, 500, 700, 1000, 1100, and 1200 hcd), and a correction rgb coefficient is similarly determined for these reference rgb coefficients, and a table composed of the reference rgb coefficient and the correction rgb coefficient for each luminous intensity. Can be created and stored in the image processing apparatus 30. When such a table is created, the luminous intensity of the headlight L may not necessarily correspond to the luminous intensity of a specific table. The desired value can be obtained by insertion.

6). Zone division In one embodiment, the correction rgb coefficient is determined for each individual pixel. However, from the viewpoint of processing efficiency, in another embodiment, all pixels of the image sensor are divided into predetermined number of zones. It is also possible to determine a common correction rgb coefficient for each zone. When zone division is performed, there are cases where all the zones have the same number of pixels and cases where at least one zone has a different number of pixels. Although any zone division can be applied in the present invention, here, a case where all the zones have the same number of pixels will be described as one embodiment.

  In the embodiment shown in FIG. 25, the image sensor 20b has a measurement region composed of 640 × 480 pixels 40 in the horizontal × vertical direction. These entire pixels are divided into zones 41 made up of 20 × 20 pixels, and the horizontal × vertical zones are divided into 32 × 24 zones 41. In each zone 41, RGB pixels 40 are arranged in a so-called buyer array. In FIG. 25, block numbers 3 to 6 are assigned to some zones 41 near the center (for example, block number 4 is assigned to the zone (12, 9)). Each block number points to a block in the table that stores the corresponding rgb coefficient. FIG. 26 shows an example of such a table of rgb coefficients. For example, in FIG. 26, the rgb coefficients indicated by the block number 5 are r = 0.4315, g = 0.5369, and b = 0.0317. Therefore, for example, since the block number 5 is assigned to the zone (12, 11) in FIG. 25, r = 0.4315, g = 0.5369, and b = 0.0317 rgb for that zone. The coefficient value is applied.

  In this way, as shown in FIG. 25, since a block number is assigned to each zone 41, a correspondence table between the zone 41 and the block number is stored in the image processing apparatus 30, Further, a correspondence table between block numbers and rgb coefficients as shown in FIG. 26 is also stored in the image processing apparatus 30. Therefore, when calculating the luminance Y (luminous intensity) for a specific pixel 40, first, it is determined which zone 41 the pixel 40 belongs to, and the corresponding block number is determined. The calculation is performed using the rgb coefficient indicated by. In FIG. 25, block numbers are assigned only to the selected zone 41 near the center instead of all zones 41, but block numbers can be assigned to all zones 41 in the same manner. As another example, it may not be necessary to use the entire measurement range of the image sensor 20b of FIG. 25, in which case, as shown in FIG. It is possible to define a smaller effective range and assign a block number only to the zone 41 belonging to the effective range.

  The examples shown in FIGS. 25 and 26 are rgb coefficients when the luminous intensity is 400 hcd, and the block number 5 corresponds to the reference rgb coefficient when 400 hcd. Therefore, a block number 5 indicating the reference rgb coefficient is assigned to many zones 41 near the center of the image sensor 20b in FIG. As shown in FIG. 26, it can be seen that the rgb coefficients indicated by the block numbers other than the block number 5 are different in value from the reference rgb coefficient, and are corrected rgb coefficients. These corrected rgb coefficients are appropriately determined by the above-described rgb coefficient correction procedure. In the block numbers 0 to 4 and 6 to 8, the corrected rgb coefficient has a coefficient value other than 0. In the block numbers 9 to 11, the corrected r coefficient and the corrected b coefficient are included in the corrected rgb coefficients. And those values are 0, and only the correction g coefficient has a value other than 0. In block number 12-14, all correction rgb coefficients are set to 0, and luminance Y is not calculated in zone 41 to which the block number is assigned.

  As described above, the correspondence table between the block numbers and rgb coefficients in FIG. 26 is obtained by adjusting the light source luminous intensity to 400 hcd. By the way, although the luminous intensity and the color temperature are not directly related, it can be said that generally, the luminous temperature tends to increase as the luminous intensity increases. Therefore, using the system shown in FIG. 23, instead of the illuminometer 58, a color temperature measuring device (a device capable of directly measuring the color temperature with a spectroradiometer or the like) is used, and the luminous intensity of the reference sphere 51 is 50 hcd. FIG. 27A shows the result of measuring the color temperature of the projection light from the reference sphere 51 when it is gradually changed to 1200 hcd. FIG. 27 (B) is obtained by fitting an example of an approximate expression to the result of FIG. 27 (A). From this result, it can be said that the color temperature of the emitted light from the light source generally increases as the luminous intensity of the light source increases. Therefore, the correspondence table between the zone 41 and the block number as shown in FIG. 25 and the correspondence table between the block number and the rgb coefficient as shown in FIG. 26 are at least the expected light intensity range of the headlight L to be measured. It is necessary to prepare for each light intensity (color temperature) and store it in the image processing apparatus 30.

  FIG. 28 shows a configuration in which all the pixels 640 × 480 in the image sensor 20b similar to FIG. 25 are equally divided into zones 41 composed of 20 × 20 pixels. In FIG. 28, a rectangular effective range 42 is defined inside a measurement range composed of 640 × 480 images of the image sensor 20 b, and block number 0 is assigned to all zones 41 outside the effective range 42. Assigned. On the other hand, block numbers 2 to 7 are assigned to the zones 41 in the effective range 42, respectively. The correspondence table between these block numbers and rgb coefficients is shown in FIG. 29. The correspondence table of FIG. 29 is the same as the correspondence table of FIG. In block number 0 assigned to zone 41 in FIG. 28, the values of their rgb coefficients are r = 0.4111, g = 0.5461, b = 0.0428, as shown in FIG. . In FIG. 29, the value of the rgb coefficient indicated by block number 5 is the reference rgb coefficient when the light source luminous intensity is 400 hcd, and therefore the block number versus rgb coefficient correspondence table of FIG. 29 is for the light source luminous intensity of 400 hcd. It is understood. Furthermore, as shown in FIG. 29, the rgb coefficients indicated by the block numbers other than the block number 5 are corrected rgb coefficients. In particular, the values of the corrected rgb coefficients indicated by the block numbers 2 to 4 and 6 to 8 are the reference values. It is understood that each value of the rgb coefficient is obtained by multiplying by a predetermined constant. The correspondence table in FIG. 29 should be applied to each zone 41 in FIG. 28, so the correspondence table between the zone 41 and block number in FIG. 29 is also for the light source luminous intensity of 400 hcd.

  As described above, by applying the corrected rgb coefficient, variations due to chromatic aberration due to the optical system (particularly, the Fresnel lens as the main lens) in the headlight tester 10 are corrected, and the luminous intensity of the headlight L is appropriately increased. It can be determined with accuracy. Furthermore, the processing can be speeded up by applying the zone division processing.

7). Correction of Maximum Brightness Position As described above, the position (h, v) of the maximum brightness Y _MAX in the light distribution pattern 33 of the traveling lamp can be determined by, for example, a 3 ° 30 minute balance method. Here, h and v are the horizontal and vertical position coordinates on the screen 19, respectively. By the way, in the system shown in FIG. 23, a reference sphere for a traveling light is used as the reference sphere 51, the reference sphere 51 is set at the selected horizontal angle position, and the projection direction is displaced stepwise in the vertical direction. When the position of the maximum luminance Y _MAX is determined by the headlight tester 10 positioned 1 m ahead of the reference sphere 51 instead of the 10 m screen 25 in the system of FIG. 23, it has been found that variations may occur. As shown in FIG. 30, for example, the reference sphere 51 is set at an angular position of 35 cm to the left (ie, h = −35), and the reference sphere 51 is spaced from the upper 10 cm (ie, v = 10) at a constant interval. Every time (in the illustrated example, every 1 cm), the angle is displaced downward to 35 cm (ie, v = −35), and the position of the maximum luminance Y _MAX is obtained using the headlight tester 10 at each angular position. Was measured. In this case, since the reference sphere 51 is merely angularly displaced from top to bottom at the same horizontal angular position, the position coordinates of (h, v) of Y _MAX obtained at the respective angular positions in the vertical direction v are the same. It is scheduled to exist on the h coordinate. However, it has been found that there is a slight variation in the headlight tester 10 in such a Y _MAX especially in the h coordinate. There is a possibility that variations occur not only in the h coordinate but also in the v coordinate. This is also considered to be due to the influence of chromatic aberration and the like due to the optical system (particularly, Fresnel lens) in the headlight tester 10.

Therefore, in order to correct such variations, as described above in the system of FIG. 23, the reference sphere 51 for the traveling lamp is used as the reference sphere 51, and the reference sphere 51 is set at each horizontal angular position to be perpendicular. The position of the maximum luminance Y _MAX is measured by the headlight tester 10 positioned 1 m ahead of the reference sphere 51 instead of the 10 m screen 25 in the system of FIG. We checked whether there was any variation. As a result, for the headlight tester 10 that requires such correction of the maximum luminance position, a correspondence table between the block numbers and the position correction values shown in FIG. 31 is created and stored in the image processing apparatus 30. doing. The correspondence table in FIG. 31 should be applied to the correspondence table between the zone 41 and the block number shown in FIG. In the embodiment shown in FIG. 31, the height position correction is zero (v = 0), but it can be seen that there are several horizontal position corrections. In another embodiment, not only horizontal position correction but also vertical position correction may be required.

  FIG. 32 shows an example of a correspondence table between zones and block numbers similar to FIG. 28, but is for correcting the position (h, v) of the balance point of the 400 hcd traveling light. FIG. 33 is a table indicating the horizontal position offset value h (cm) and the vertical position offset value v (cm) assigned to each block number in FIG. Therefore, for example, block number 5 is assigned to the zone of the horizontal pixel range (300-319) and vertical pixel range (120-139) in FIG. 32, and position correction corresponding to this block number 5 is performed. The quantities are a horizontal position offset value h = 0 and a vertical position offset v = 0. That is, when block number 5 is assigned to the zone of FIG. 32, position correction is not performed for 20 × 20 pixels belonging to that zone. In the embodiment of FIG. 32, block number 5 is assigned to most of the zones within the effective range, and only block numbers 4 and 6 are assigned to the remaining zones. It is understood that the amount of is relatively small.

8). Correction of Elbow Point Position In the system shown in FIG. 23, a horizontal (left / right) angle at which a reference sphere 51 is used as a reference sphere 51 and the reference sphere 51 is selected as shown in FIGS. Set to a position (shown as the corresponding horizontal position h (cm) on the 10 m screen in FIGS. 24 and 30), while maintaining the selected horizontal angular position, the reference sphere 51 is vertical (up and down) As shown in FIG. 23, the angle (in FIG. 24 and 30, shown as the corresponding vertical position v (cm) on the 10 m screen) is changed stepwise (in FIG. 30, every 1 cm). The position of the elbow point at each vertical position v was measured using the present headlight tester 10 positioned 1 m ahead of the reference sphere 51 instead of the 10 m screen. As the selected horizontal position h, left 35 cm, left 20 cm, left 10 cm, center (left and right 0), right 10 cm, right 20 cm, and right 35 cm were used. The results thus obtained are shown as a graph in FIG.

  Incidentally, the reference sphere mounting device 50 in the system shown in FIG. 23 is configured with extremely high accuracy, and the angular displacement of the reference sphere 51 in the horizontal and vertical directions can be performed with extremely high accuracy. The reference sphere for the passing lamp projects a light distribution pattern having a horizontal cut line, an oblique cut line, and an elbow point as an intersection of the cut lines. Accordingly, when a reference sphere for passing light is attached to the reference sphere mounting device 50 as the reference sphere 51 and the reference sphere is changed in a stepwise manner in the vertical direction at the selected horizontal angular position, the horizontal angular position A sequence of elbow points whose positions change in a straight line in the vertical direction should be obtained.

  However, as shown in the graph of FIG. 34, the measurement result was obtained when the reference sphere 51 was angularly displaced stepwise in the vertical (up and down) direction at any of the horizontal positions h. The horizontal position h of the elbow point has left and right variations. In FIG. 35, the reference sphere 51 is set at a horizontal angular position h of 10 cm (ie, −10 cm) left on the 10 m screen, and the reference sphere 51 is moved downward from the upper 20 cm (ie, v = 20) at regular intervals. This headlight tester in which the angle (h, v) of the elbow point is located 1 m ahead of the reference sphere 51 by shifting the angle stepwise and linearly down to 35 cm downward (ie, v = −35). 10 is a graph showing a positional deviation in the vertical (up and down) direction and a horizontal (left and right) direction when measured at 10. FIG. From this graph, it can be seen that the position of the elbow point measured by the headlight tester 10 varies not only in the horizontal (left and right) direction but also in the vertical (up and down) direction. The horizontal (left-right) direction position h of the measured elbow point should originally be kept constant at a position of h = −10 cm, but varies between h = −10 and h = −20. doing. Further, it is shown that the vertical (up and down) position v of the measured elbow point is not aligned on one straight line, and there is slight slight variation.

  The reason why such variation occurs when the h and v positions of the elbow point are measured using the headlight tester 10 is that an optical system is used to establish 10 m reproduction in the headlight tester 10. Therefore, it seems to be due to the influence of chromatic aberration and the like. In particular, when a Fresnel lens is used as the main lens 17 of the optical system, the influence is considered to be more conspicuous due to the structural speciality of the Fresnel lens.

  Therefore, when measuring the positions h and v of the elbow point in the headlight tester 10, it is important to correct the above-described horizontal and vertical variations. Therefore, in the present invention, the reference sphere for passing lamp is set as the reference sphere 51 using the system shown in FIG. 23, and the reference sphere 51 is set at the selected angular position h in the horizontal direction. 51, the angular position v is changed stepwise in the vertical direction to determine the position (h, v) where the elbow point should be on the 10 m screen 25 from the angle scale of the reference ball mounting device 50, while in the same setting state The position correction amount is determined when there is a difference between the elbow point position (h, v) measured by the headlight tester 10. The position correction tables thus determined are shown as (A) to (C) in FIG. 36A shows the correspondence between the block number and the horizontal correction, FIG. 36B shows the correspondence between the block number and the cut line correction, and FIG. 36C shows the block number and the height correction. The correspondence with is shown. As shown schematically in FIG. 37, the horizontal correction (A) in FIG. 36 is a horizontal position correction h and a vertical position correction v with respect to the measurement position of the horizontal cut line, and (B) The cut line correction of (C) is the horizontal position correction h and the vertical position correction v with respect to the measurement position of the oblique cut line, and the height correction of (C) is the second above the horizontal cut line. The horizontal position correction h and the vertical position correction v with respect to the measurement position of the horizontal cut line. Note that the height correction (C) is actually a correction amount applied when the light distribution of the headlight L is a Z-beam distribution. A correspondence table (not shown) indicating to which pixel 40 (or zone 41) of the image sensor 20b the block number shown in each correction table shown in FIG. 36 is also shown, for example, in FIG. In this form, it is prepared and stored in the image processing apparatus 30.

9. Search for Bright / Dark Branch Line (1) Emphasis on Dark Field FIG. 38 is a visual sensitivity diagram showing the sensitivity of the human eye. Bright field sensitivity with the highest sensitivity at 555 nm and dark field sensitivity with the highest sensitivity at 507 nm It is well known that there is. In an original headlight test, an inspector looks for a light-dark branch line by visual observation of a light distribution pattern projected on a 10 m screen located 10 m ahead of the headlight in a dark room. In that case, it is considered that the inspector searches for the bright and dark branch lines with dark field sensitivity rather than bright field. Therefore, in order to realize such a situation in the headlight tester 10 as well, when searching for a light / dark branch line in the light distribution pattern, the RGB data obtained by the color camera 20 for measurement from the light distribution pattern is used. The luminance Y may be determined by relatively enhancing the B data as compared with the R data and the B data.

  For example, when a light distribution pattern of projection light from an HID headlamp (color temperature = 4500 K) is imaged by the measurement color camera 20 of the headlight tester 10, the following digital RGB data is output from the measurement color camera 20. Is output. The default color temperature of the measurement color camera is 3140K.

R = 181
G = 230
B = 240
Since the reference rgb coefficients for the color temperature 3140K (determined based on the color temperature of the CIE standard light source) are r = 0.4111, g = 0.5461, b = 0.0428, these reference rgb coefficients are used. The luminance Y is calculated as follows.

Y = 0.4111 × 181 + 0.5461 × 230 + 0.0428 × 240
= 210
On the other hand, since the reference rgb coefficients for the color temperature of 4500 K are r = 0.37186, g = 0.55918, and b = 0.0896, the luminance Y is calculated using these reference rgb coefficients as follows. is there.

Y = 0.37186 × 181 + 0.55918 × 230 + 0.06896 × 240
= 212
These differences indicate that the influence of B becomes stronger as the color temperature increases. In one embodiment of the present invention, the influence of B is relatively strengthened in RGB data. For example, in the above example, r = 0.2, g = 0.6, b = 0. Determine the rgb coefficient as in .2. The luminance in that case is calculated as follows.

Y = 0.2 × 181 + 0.6 × 230 + 0.2 × 240
= 222
In this way, a luminance Y with a relatively large influence of B is obtained. The plurality of luminances Y thus obtained can be compared with each other by, for example, the difference difference method shown in FIG. The bright and dark branch lines obtained in this way are close to those obtained by human visual observation (dark field sensitivity). In the above example, r + g + b = 1 and the Y coefficient 1 is maintained. However, by increasing the ratio of the b coefficient relatively between the r coefficient, the g coefficient, and the b coefficient, the influence of the blue component is increased in luminance. Y is strongly reflected. The special rgb coefficient for searching for the light / dark branch line determined in this way is stored in the image processing apparatus 30 and is used only when the light / dark branch line in the light distribution pattern 33 is searched. Used in. In another example, the ratio of the b coefficient is increased relative to the r coefficient and the g coefficient without maintaining the Y coefficient 1.

(2) The correction coefficient is only the g coefficient. In this embodiment, the reference rgb coefficient (determined by the color temperature of the CIE standard light source) and the correction rgb coefficient are used. Among the correction rgb coefficients, the r coefficient and the b coefficient are used. Is 0, and correction is performed only by the correction g coefficient. One example is shown in FIG. The example of FIG. 39 is for a luminous intensity of 400 hcd, and a similar correspondence table is created for the other luminous intensity and is stored in the image processing apparatus 33 together. As shown in the graphs of FIGS. 21 and 22, G is the most stable among the RGB data. In particular, the RGB data is relatively stable near the center of the light distribution pattern, but the fluctuations in R and B increase toward the periphery. Accordingly, the reference rgb coefficient is applied in the vicinity of the center, but in the peripheral part, the r coefficient and the b coefficient are set to 0, and only the g coefficient is set as an effective correction coefficient. That is, using the system shown in FIG. 23, a low-light reference sphere is attached as a reference sphere 51, and its luminous intensity is measured with an illuminometer 58 while angularly displacing it up and down and left and right. The correction g coefficient is determined so that the same light intensity value can be obtained by comparing with the result of the same measurement in the headlight tester 10. The special rgb coefficient for searching for the light / dark branch line determined in this way is stored in the image processing apparatus 30 and is used only when the light / dark branch line in the light distribution pattern 33 is searched. Used in.

  The block numbers shown in the correspondence table of FIG. 39 can be applied to, for example, FIG. 25 showing the correspondence table between the block numbers and the zones 41 for the same 400 hcd luminous intensity. As shown in the correspondence table of FIG. 39, in this case, the condition of Y coefficient 1, that is, r + g + b = 1 is violated.

(3) Proportional Setting by Multiplier α FIG. 40A shows a light distribution pattern 33 of the projection light from the passing headlight L captured by the color camera for measurement 20, and the condition of the Y coefficient of 1, that is, r + g + b = 1. A state in which the luminance Y is calculated using a satisfactory rgb coefficient is shown. For the light distribution pattern 33 of FIG. 40 (A), the headlight tester 10 calculates the luminance Y using the rgb coefficient satisfying the condition of Y coefficient 1, ie, r + g + b = 1, and thus obtained. As a result of searching for a bright and dark branch line by applying a difference difference = 0 to a plurality of luminances Y, it has been found that the rise of the cut line may sink as shown in FIG. . This is considered to be because the cut line as the boundary between the dark part and the bright part in the light distribution pattern 33 is not so clear and is dragged toward the bright part because the brightness at the center of the bright part is high. . Thus, it has been found that the bright and dark branch lines may not be searched accurately if the Y coefficient is 1 as it is.

  Therefore, regardless of the theoretical condition of the Y coefficient 1, the reference rgb coefficient is multiplied by a multiplier α to determine a special correction rgb coefficient for light and dark branch line search, and these correction rgb coefficients are used. When the luminance Y was calculated and the light / dark branch line was searched with the difference difference = 0, it was found that the light / dark branch line could be searched quickly and accurately. As a result of various trials and errors, it has been found that it is appropriate that the value of the multiplier α is greater than 1 and 4 or less. For example, in FIG. 41A, the light distribution pattern 33 of the light projected from the passing headlight L is imaged by the measurement color camera 20, and the rgb coefficient satisfying the condition that the Y coefficient is 4, that is, α = 4 is used. The brightness Y is calculated. α = 4 means that each rgb coefficient is multiplied by a multiplier α such as α × r, α × g, α × b. When a light / dark branch line is searched by applying a difference difference = 0 to a plurality of luminances Y thus obtained, as shown in FIG. Turned out to be. When the multiplier α is increased beyond 4, the light and dark branch line obtained by processing with the difference difference = 0 is floated above the actual cut line, and the value of the multiplier α is too large. I mean.

  Therefore, when a multiplier α of a desired value is stored in the image processing device 30 and the headlight tester 10 searches for the light / dark branch line in the light distribution pattern 33 of the passing lamp, the multiplier α is applied to each coefficient of rgb. To determine a special rgb coefficient to be used for the search for the bright and dark branch lines.

(4) Gamma correction The above three methods are for searching for a light / dark branch line by correcting the rgb coefficient in the light / dark branch line search mode. In this embodiment, the light distribution pattern image is corrected by gamma correction. In this case, a special rgb coefficient for searching for a bright and dark branch line is not used. That is, when gamma correction is applied in the headlight tester 10, there are a case where it is applied to RGB data itself as color image data and a case where it is applied to luminance Y as grayscale data.

  When applied to RGB data itself, it is as follows.

R _OUT = 255 × (R _IN / 255) ^{1 / γ}
G _OUT = 255 × (G _IN / 255) ^{1 / γ}
B _OUT = 255 × (B _IN / 255) ^{1 / γ}
Note that γ = 1.5. R _IN , G _{IN, and} B _IN are RGB data before gamma correction processing, and R _OUT , G _OUT , and B _OUT are RGB data after gamma processing. In this way, it is possible to calculate luminance Y using RGB data to which gamma correction is applied, and to search for a bright and dark branch line by applying a difference difference = 0 to the luminance Y.

  On the other hand, when applied to the luminance Y, it is as follows.

Y _OUT = 255 × (Y _IN / 255) ^{1 / γ}
Note that γ = 1.5. Y _IN is the luminance Y calculated using a predetermined rgb coefficient, and Y _OUT is a result of performing gamma correction processing on Y _IN . Therefore, it is possible to search the dark dividing lines by applying the difference = 0 of the difference with respect to Y _OUT.

  Performing such a gamma correction and searching for bright and dark dividing lines is particularly effective when the headlight L is a halogen lamp.

(5) Natural logarithmic correction In this case as well, there are a case where it is applied to RGB data itself which is color data and a case where it is applied to luminance Y which is grayscale data. First, when applied to RGB data, it is as follows.

R _OUT = 255 × (ln R _IN / ln 255)
G _OUT = 255 × (ln G _IN / ln 255)
B _OUT = 255 × (ln B _IN / ln 255)
Note that R _IN , G _{IN, and} B _IN are RGB data before the natural logarithmic correction process, and R _OUT , G _OUT , and B _OUT are RGB data after the natural logarithmic correction process.

In this way, it is possible to calculate luminance Y using RGB data to which natural logarithmic correction is applied, and to search for a bright and dark branch line by applying a difference difference = 0 to the luminance Y.

  On the other hand, when applied to the luminance Y, it is as follows.

Y _OUT = 255 × (ln Y _IN / ln 255)
Y _IN is the luminance Y calculated using a predetermined rgb coefficient, and Y _OUT is a value obtained by performing natural logarithmic correction processing on Y _IN . Therefore, it is possible to search the dark dividing lines by applying the difference = 0 of the difference with respect to Y _OUT.

  Performing such a natural logarithmic correction and searching for bright and dark dividing lines is particularly effective when the headlight L is a lamp other than a halogen lamp.

10. Measurement Principle and Procedure According to the principle of the present invention, the rgb coefficient is determined based on the color temperature of the CIE standard light source, and RGB data is determined by color image processing the projection light of the headlight, and the RGB data and the rgb The luminance Y is determined using the coefficient. The color temperature expresses the color of the light source and uses the temperature of a black body that appears to be the same color as that color, and is not an index indicating the brightness of the light source. In the present invention, since the headlight is also a light source, the rgb coefficient used when the projection light of the headlight is imaged by a color camera and the light distribution pattern is processed in a color image depends on the color temperature of the headlight. It focuses on what should be decided. If the color temperature of the headlight can be directly determined by any method, an rgb coefficient corresponding to the color temperature may be determined (selected). On the other hand, if the headlight color temperature cannot be determined directly, the headlight brightness (color temperature) is estimated, and the estimated brightness (color temperature) can be brought close to the true color temperature of the headlight. It ’s fine. For example, a temporary rgb coefficient is determined using the estimated light intensity (color temperature), a temporary brightness is determined using these temporary rgb coefficients, and a temporary light intensity (color temperature) is determined from the temporary brightness. And the provisional luminous intensity (color temperature) is made closer to the true color temperature of the headlight than the estimated color temperature. Then, if the temporary luminous intensity (color temperature) exceeds the tolerance from the initially estimated luminous intensity (color temperature), a new rgb coefficient corresponding to the temporary luminous intensity (color temperature) is determined, Thereafter, the same processing may be performed.

Then, by searching for the maximum brightness Y _MAX among the brightness Y determined as described above, the position of the maximum brightness Y _MAX can be determined as the optical axis (irradiation direction) of the headlight. This corresponds to determining the irradiation direction of the projection light when the headlight is in the traveling light mode. Furthermore, the maximum luminance Y _MAX can be converted into luminous intensity (hcd) according to a predetermined luminance / luminous intensity conversion formula. This corresponds to determining the luminous intensity (hcd) of the headlight when the headlight is in the traveling light mode.

  On the other hand, it is possible to determine the luminance Y corresponding to a predetermined position (such as a road surface irradiation point) from the luminance Y determined as described above, and convert the luminance Y into luminous intensity. This corresponds to determining the light intensity at the road surface irradiation point when the headlight is in the passing mode.

  Further, the luminance Y determined as described above is compared with each other to find a horizontal cut line and a diagonal cut line as light and dark dividing lines, and an elbow point can be determined as an intersection of these horizontal and diagonal cut lines. It is.

  When searching for a bright and dark branch line by comparing the luminance Y obtained from the light distribution pattern with each other, the luminance Y is calculated using a special rgb coefficient for searching for the bright and dark dividing lines. It is possible to search for light and dark branch lines by applying the difference difference = 0. In one embodiment, such a special rgb coefficient increases the b coefficient relative to the other coefficients to enhance the blue component and make it closer to the visual search for bright and dark dividing lines. In another embodiment, it is possible to set the r and b coefficients to 0 and calculate the luminance Y as a special rgb coefficient using only the g coefficient set against the luminometer. . In yet another embodiment, the rgb coefficient is multiplied by the same multiplier α (1 <α ≦ 4) to determine a special rgb coefficient, and the luminance Y is calculated using the special rgb coefficient. Is possible. In yet another embodiment, gamma correction or natural logarithm correction is performed on the RGB data or luminance Y without using a special rgb coefficient, and the resulting luminance Y is obtained by applying a light-dark branch line. It is possible to perform a search.

  Next, several embodiments of the procedure for testing the headlight in the present invention will be described.

  First, the headlight and the headlight tester are relatively moved to a position where the light distribution pattern of the projection light from the headlight can be imaged by the color camera for measurement of the headlight tester. The image processing apparatus of the headlight tester extracts the maximum luminance in the light distribution pattern and adjusts the shutter speed of the measurement color camera by increasing or decreasing the shutter speed so that the maximum luminance falls within a predetermined range. When the shutter speed is set to an appropriate value, the image processing apparatus determines the light source type of the headlight using the measurement color camera. Typically, the light source type includes a halogen lamp, an HID lamp, and an LED lamp. Halogen lamps are also referred to as warm color lamps, while HID lamps are also referred to as cold color lamps. In particular, the color temperature differs between the warm color lamp and the cold color lamp, and a basic rgb coefficient for the warm color lamp and a basic rgb coefficient for the cold color lamp are prepared. The basic rgb coefficient is a representative and basic rgb coefficient in the light source type. For example, in the case of a warm color light source type, the rgb coefficient is a rgb coefficient with respect to a luminous intensity of 400 hcd. Therefore, by determining the light source type, the basic rgb coefficient of the light source type can be used as a temporary rgb coefficient even if the color temperature of the headlight cannot be directly determined. . For example, if the headlight to be tested is a warm color lamp, the luminous intensity of the headlight is provisionally set to 400 hcd, which is estimated by estimating the color temperature of the headlight to a color temperature corresponding to 400 hcd. Is the same as The basic rgb coefficient is also determined in advance for the cold-colored lamp and the LED lamp and stored in the image processing apparatus.

  After discriminating the light source type, the lamp type of the headlight (projector, lens cut, multi-reflector) is performed. This is preferably performed by a facing camera. Further, the image processing apparatus uses the facing camera to align the center of the headlight with the optical axis of the headlight tester. This operation is usually called a right-to-back operation.

  When the facing is completed, the process is divided depending on the measurement mode, which is the traveling light mode or the passing light mode. In the traveling light mode, there are two embodiments, a direct method and an iterative method, as follows, depending on whether or not the color temperature of the headlight can be determined directly.

A. Traveling Light Mode (1) Direct Method First, when the color temperature of the headlight L can be determined directly, the direct method is adopted. This is because, if the color temperature of the headlight L is determined, the rgb coefficient to be applied to the projection light from the headlight L can be directly determined (selected). For example, in one embodiment, the headlight tester 10 is provided with a color temperature measuring device, and the projection light from the headlight L to be tested is detected to determine the color temperature of the headlight. Information on the color temperature is supplied to the image processing apparatus 30. In another embodiment, the operator operates a commercially available color temperature measuring device to measure the color temperature of the headlight, and the operator inputs the color temperature information to the headlight tester 10. The measurement color camera 20 captures the light distribution pattern 33 of the projection light from the headlight L and supplies RGB data to the image processing device 30. The image processing device 30 uses the RGB data and the rgb coefficient to produce a color image. The luminance Y for the pixel 40 of the sensor 20b is calculated.

  Note that, due to the influence of the optical system (particularly, Fresnel lens) used by the headlight tester 10, there may be a case where the variation in luminous intensity based on the variation in RGB data cannot be ignored within the measurement range of the color image sensor 20b. In such a case, the rgb coefficient determined based on the color temperature of the CIE standard light source is used as the reference rgb coefficient, and the position within the measurement range to which the reference rgb coefficient is applied is specified, while such variations are absorbed. Then, a corrected rgb coefficient obtained by correcting the reference rgb coefficient is determined, and a position within the measurement range to which the corrected rgb coefficient is applied is specified. These reference rgb coefficients, corrected rgb coefficients, and a correspondence table between these coefficients and application positions in the measurement range are stored in the image processing apparatus 30. Further, in one embodiment, such a correspondence table is created for all the pixels 40 within the measurement range. However, in another embodiment, all the pixels 40 within the measurement range are assigned to a plurality of pixels. The zone 41 is composed of pixels 40 and a common correspondence table is applied to each zone 41. Each zone 41 may have the same number of pixels or different numbers of pixels.

The image processing apparatus 30 searches for the maximum brightness Y _MAX among the brightness Y calculated as described above. Next, the image processing apparatus 30 determines the position of the maximum luminance Y _MAX as the optical axis (irradiation direction) of the headlight L. Further, the image processing apparatus 30 converts the maximum luminance Y _MAX into the luminous intensity (hcd) according to a conversion formula between the predetermined luminance Y (dimensionless) and luminous intensity (hcd). In this way, the optical axis (irradiation direction) and luminous intensity (hcd) of the headlight L are determined by the direct method.

(2) Iterative method Next, when it is impossible to directly determine the color temperature of the headlight L, an iterative method is adopted. In this case, the basic rgb coefficient specified for each light source type in advance is used as a temporary rgb coefficient using the above-described determination result of the light source type. For example, when the determination of the light source type is a warm color system (halogen lamp), the image processing apparatus uses an rgb coefficient of 400 hcd as the temporary rgb coefficient as the temporary rgb coefficient. The measurement color camera 20 captures the light distribution pattern 33 of the projection light from the headlight L and supplies RGB data to the image processing device 30. The image processing device 30 uses the RGB data and the provisional rgb coefficient. A temporary luminance Y for the pixel 40 of the color image sensor 20b is calculated.

  Note that, due to the influence of the optical system (particularly, Fresnel lens) used by the headlight tester 10, there may be a case where the variation in luminous intensity based on the variation in RGB data cannot be ignored within the measurement range of the color image sensor 20b. In such a case, the basic rgb coefficient is set as a reference rgb coefficient, a position within the measurement range to which the reference rgb coefficient is applied is specified, and a correction in which the reference rgb coefficient is corrected so as to absorb such variations. The rgb coefficient is determined, and the position within the measurement range to which the corrected rgb coefficient is applied is specified. These reference rgb coefficients, corrected rgb coefficients, and a correspondence table between these coefficients and application positions in the measurement range are stored in the image processing apparatus 30. Further, in one embodiment, such a correspondence table is created for all the pixels 40 within the measurement range. However, in another embodiment, all the pixels 40 within the measurement range are assigned to a plurality of pixels. The zone 41 is composed of pixels 40 and a common correspondence table is applied to each zone 41. Each zone 41 may have the same number of pixels or different numbers of pixels.

The image processing device 30 searches for the temporary maximum luminance Y _MAX among the temporary luminances Y calculated as described above. Next, the image processing apparatus 30 converts the temporary maximum luminance Y _MAX into the temporary luminous intensity (hcd) according to a conversion formula between a predetermined luminance Y (dimensionless) and luminous intensity (hcd).

Next, when the provisional luminous intensity obtained in this way exceeds the allowable difference and is different from 400 hcd estimated based on the light source type determination result, the image processing device 30 is determined in this way. A new rgb coefficient (preferably, a combination of the reference rgb coefficient and the correction rgb coefficient) is selected based on the temporary luminous intensity (hcd), and the RGB data of the light distribution pattern 33 supplied from the measurement color camera 20 A new luminance Y for the pixel 40 of the color image sensor 20b is calculated using the new rgb coefficient. The image processing device 30 detects the maximum new maximum luminance Y _MAX among these new luminances, and further calculates a new maximum luminance Y _MAX according to the conversion formula, preferably the final luminous intensity. To (hcd). If necessary, it is possible to select a new rgb coefficient again based on the new luminous intensity and repeat the process until a final luminous intensity having a desired accuracy is obtained or a predetermined number of repetitions. Then, the image processing apparatus 30 determines the optical axis (irradiation direction) of the headlight L by applying the maximum luminance (for example, balance point) position correction table to the position of the maximum luminance Y _MAX with respect to the final luminous intensity.

B. Passing light mode When the passing light mode is selected, either (1) direct method or (2) repetition method described above can be applied to determine the luminous intensity (hcd). It may not be necessary to use a corrected rgb coefficient because the variation in data is substantially negligible. In the low-light mode, the place where the light intensity is measured is not the highest luminance point but a point called a road surface irradiation point defined in advance with respect to the lamp center. That is, when the lamp height is 1 m or less, the road surface irradiation point is a position 1.3 degrees to the left and 0.6 degrees below the lamp center, and when the lamp height is more than 1 m, it is left from the lamp center. The position is 1.3 degrees and 0.9 degrees below. Therefore, the luminance Y may be calculated from the RGB data and the rgb coefficient at these fixed positions and converted into luminous intensity (hcd) according to the above conversion formula.

  On the other hand, in the passing light mode, it is necessary to detect the position of the elbow point. Therefore, the image processing apparatus 30 calculates the luminance Y for each pixel in the measurement range of the color image sensor 20b from the rgb coefficient and the RGB data, and compares them with each other to perform horizontal cutting as a light / dark branch line. Find lines and diagonal cut lines. One embodiment for performing the comparison process is a difference difference method. Further, since the influence of the optical system (Fresnel lens 17) of the headlight tester 10 is relatively great in detecting the position of the elbow point, a position correction table is applied to correct the position of the detected elbow point.

  When detecting the elbow point, first, in order to determine the cut line, the method of the present invention enters the light / dark branch line search mode, and the light / dark branch line that is the boundary between the bright part and the dark part in the light distribution pattern. Search for. When searching for a light / dark branch line, in one embodiment, the rgb coefficient is converted into a special rgb coefficient for light / dark branch line search, and the luminance Y in the light distribution pattern is calculated using the special rgb coefficient. Then, the bright and dark branch lines are searched by comparing their luminances. In that case, it is possible to search for the bright and dark dividing lines by applying the difference difference = 0 to the luminance Y. Further, in another embodiment, gamma correction or natural logarithmic correction is performed on RGB data or luminance Y without converting the rgb coefficient into a special rgb coefficient for searching for a bright and dark dividing line, and such correction is performed. The brightness obtained from the RGB data or the corrected brightness are compared with each other to search for a light / dark branch line. In that case, it is possible to search for a bright and dark branch line by applying the difference difference = 0 to the luminance Y.

  As described above, various embodiments when a color camera is used as the measurement camera 20 in the present invention have been described in detail. Next, a configuration and a method for facing the headlight tester in the present invention will be described in detail. In the present invention, either the color camera or the monochrome camera can be used as the facing camera 24. In one preferred embodiment, the monochrome camera equipped with the CMOS image sensor is used. Use the camera.

11. Face-to-face (1) Mutual use of information of measurement camera and face-to-face camera In one embodiment of this aspect of the invention, headlight tester 10 includes not only measurement camera 20 but also face-to-face camera 24. It has. These cameras are connected to the image processing device 30, and the measurement camera 20 supplies information regarding the light distribution pattern of the headlight L projected on the measurement screen 19, while the facing camera 24 is a head. An image of the light L is taken and information about the image is supplied. Therefore, the image processing apparatus 30 can perform comprehensively improved image processing by mutually utilizing the light distribution pattern information and the headlight image information. For example, when the headlight tester 10 finishes the facing operation, the optical axis of the headlight tester (the center 17a of the main lens 17) deviates from the lamp center (for example, the bulb center) of the headlight L within a predetermined allowable range. In such a case, it is possible for the image processing device 30 to identify such a positional deviation from the headlight image information captured by the facing camera 24. The image processing apparatus 30 can improve the measurement accuracy by using such misregistration information as a correction value when processing information regarding the light distribution pattern 33 captured by the measurement camera 20. It is.

  Furthermore, the image processing apparatus 30 can specify that the headlight L is a projector lamp when the image is circular by receiving the headlight image information from the facing camera 24. It is. Further, the halogen lamp has a lens cut and a multi-reflector, while the HID lamp has a projector and a multi-reflector, and the lens cut and the multi-reflector are prominent in the size and pitch of a plurality of vertical stripes in the lamp image. There are significant differences. Therefore, the image processing apparatus 30 processes the light distribution pattern information from the measurement camera 20 to determine whether the lamp is a halogen lamp or an HID lamp based on the difference and ratio between the two pieces of RGB data as described above. Based on this identification information and the vertical stripe information in the headlight image from the facing camera 24, it is possible to identify whether it is a multi-reflector or a lens cut.

  Further, when the image processing device 30 has an almost circular image shape based on the headlight image from the facing camera 24, it is possible to specify that the headlight L is a projector, When the image processing apparatus 30 processes the light distribution pattern information from the measurement camera 20, it is possible to estimate how the light distribution of the projector lamp can be made, and the angle of the oblique cut line in the light / dark branch Can be estimated, and the process of specifying the light-dark branch line is facilitated.

  In this way, by mutually using the light distribution pattern information from the measurement camera 20 and the headlight image information from the facing camera 24, the image processing apparatus 30 can perform a headlight test quickly and with high accuracy. It becomes possible to carry out.

(2) Lamp center (center of facing)
As a result of examining the images of various types of headlights L taken with the facing camera 24, the center of the lamp of the headlight L is around the headlight L regardless of the type or type of the headlight L. On the other hand, it has been found that there is a dark part having a certain characteristic. One example thereof is shown in FIG. 42, which are captured images of various headlights captured using the HDR function by a direct-facing camera (monochrome CMOS image sensor) 24, and (A) is a lens cut. , (B) is a multi-reflector, (C) is another multi-reflector, (D) is a projector, and (E) is another multi-reflector. 42. The circles having the same radius existing in each captured image in FIG. 42 are included in the captured image as reference circles. For example, the image processing apparatus 30 refers to the reference circle (D ) Can be used when determining the projector lamp. In order to capture such a lamp image of the headlight L with the front-facing camera 24, it is necessary to have a relatively wide dynamic range so as not to cause overexposure, blackout, and the like. In the embodiment, imaging is performed using a high dynamic range (HDR) function provided in the facing camera 24. The reason why the dark portion exists in the center of the lamp is that a bulb called “cap” is attached to the bulb in the lamp of the headlight L, or the tip portion of the bulb itself is colored black. Furthermore, even if the light from the bulb is not directly shielded in this way, for example, the reflected light from the back is bright as in the multi-reflector type, so that the vicinity of the bulb center is relative. Because it will appear dark. For example, it can be understood that the thickness and pitch of a plurality of vertical stripes in the lamp image are clearly different between (A) the lens cut and (B) the multi-reflector in FIG. By applying, the image processing apparatus 30 can identify them.

  Thus, since there is a characteristic dark part at the center of the lamp of most headlights L regardless of the lamp type or type, if such dark part can be appropriately image-processed, the headlight L The center of the lamp (the center of the bulb) can be determined as the center of facing. Various image processing technologies were examined from such a viewpoint, and a plurality of morphological calculation processes, particularly a closing process (a process in which erosion is performed the same number of times as expansion after performing expansion (dilation)). It was found that the dark part at the center of the lamp can be shaped into a predetermined shaped image by applying the lamp twice, and the center of the shaped image can be specified as the center of the lamp as the center of the lamp (bulb center). .

  Next, a process for obtaining a desired shaped image by applying such a closing process a plurality of times will be described based on one embodiment of the present invention.

  FIG. 43 shows a lamp image of the headlight L captured by the facing camera 24. Next, FIG. 44 shows a binarized image obtained by performing binarization processing on the lamp image of FIG. Next, FIG. 45 shows the result of applying the first closing process to the binarized image of FIG. FIG. 46 to FIG. 57 show the results of the sequential closing process, and FIG. 57 shows the result of the thirteenth closing process. As shown in FIG. 57, in the present embodiment, it can be seen that a substantially rectangular shaped image is formed corresponding to the dark portion existing near the center of the lamp. Accordingly, the image processing apparatus 30 can determine the center of the rectangular shaped image, and can set the center determined in this way as the directly-facing center.

  The shape of the shaped image finally obtained in this way may depend on the structure of the structuring element used in the closing process, and therefore the predetermined shape is not only the above-described rectangular shape but also a square or other shape. It is also possible to have a shape of Further, in the above-described embodiment, the closing process is repeatedly applied 13 times. However, the closing process may be repeated until a shaped image having a predetermined shape or a predetermined size is obtained. In the above-described embodiments, the closing process is used as the morphological operation process, but it is needless to say that other morphological processes can be applied.

  As described above, according to this embodiment of the present invention, the image processing device 30 uniformly uses the lamp center (bulb center) of the headlight L as the center of facing regardless of the lamp type and type of the headlight L. It is possible to determine. Next, the image processing device 30 supplies a movement control signal to the up / down / left / right driving device and the detector 14d via the control unit 31, and the headlight tester 10 is moved relative to the headlight L to be in a facing state. The

(3) Misalignment and correction thereof The detailed configuration for controlling the movement of the headlight tester 10 in the vertical and horizontal directions is not shown, but such a configuration is usually configured by a drive source such as a motor or a gear. Alternatively, it includes a mechanical power transmission mechanism such as a belt and a position control detector such as a limiter. In such a configuration, since there is a clearance between the elements, in order to perform the movement control of the headlight tester 10 with high accuracy, these various elements may be highly accurate. Further, in order to position the headlight tester 10 in a directly-facing state, the headlight tester 10 may need to be moved by reciprocating up and down or left and right several times, and the operation is delayed. There is also a case.

  Therefore, in this embodiment of the present invention, when the headlight tester 10 is moved with respect to the headlight L to perform the facing, the position of the lamp center (bulb center) 60a of the headlight L and the headlight tester 10 When the center of the optical axis (main lens center 17a) falls within a predetermined allowable range, the centers are substantially aligned and the alignment is substantially completed. In this case, a positional deviation has occurred between the position of the lamp center 60a and the position of the optical axis center of the headlight tester 10, but it is ensured that it is within an allowable range. The positional deviation information can be determined by the image processing apparatus 30. Therefore, the image processing apparatus 30 can appropriately perform correction in the subsequent processing on the assumption that such a positional deviation exists. In this case, since the headlight tester 10 does not need to be exactly aligned with the directly-facing position, the system configuration and processing are simplified and the operation is speeded up.

  That is, in the embodiment shown in FIG. 58, assuming that the positional deviation of 10 mm is within an allowable range, the headlight tester with respect to the lamp center 60a of the headlight L is completed when the alignment is completed. 10, that is, the optical axis of the light receiving unit 11 (center 17a of the Fresnel lens 17) is shifted to the right by 10 mm. Therefore, the measurement camera 20 attached to the light receiving unit 11 is displaced by 10 mm to the right with respect to the lamp center 60a of the headlight L. Therefore, the light distribution pattern 33 projected on the measurement screen 19 is also shifted to the right by 10 mm with respect to the lamp center 60a. In FIG. 58, the lamp image 60 captured by the facing camera 24 and the light distribution pattern 33 projected on the measurement screen are each rotated 90 degrees forward for easy viewing. On the other hand, the intersection of the horizontal line H and the vertical line V of the 10 m screen 25 is aligned with the lamp center 60a of the headlight L, and there is no positional deviation in the light distribution pattern 33 on the 10 m screen 25.

  FIG. 59 shows a displacement state of the lamp image 60 by the facing camera 24 during facing. That is, the facing operation of the embodiment in this case will be described below. The headlight tester 10 receives a measurement start command from the storage position, starts operation, moves to the measurement position of the left or right headlight L according to a preset measurement order, and is provided with a separately provided traversing encoder (not provided). It stops by setting a stop position by distance setting or the like or by engaging with a limit switch (not shown).

  If the lamp image is reflected in the field of view of the facing camera 24, the image processing device 30 measures the lamp center measured by the method set from there to the facing setting position in the preset field of view. The light receiving unit 11 is moved vertically and horizontally to a position where the positions overlap. If the lamp image is not reflected within the field of view of the facing camera 24, the manual button is pressed to enter the manual mode, and the operator manually moves to the position where the lamp image is captured by the facing camera 24. And track.

  When reaching the directly-facing setting position, the image processing device 30 calculates the difference in relative position from the top, bottom, left, and right 0-0 positions corresponding to the center 17a of the Fresnel lens 17 of the light receiving unit 11, and the light receiving unit corresponding to the pixel number 11 is moved to the Fresnel lens center 17a (referred to as a predictive direct control mode). When the light receiving unit 11 stops in the vicinity of the lamp center 60a in the predictive facing control mode, the image processing apparatus 30 again measures the position of the center 60a of the lamp image 60 again by the same method. In the case where there is an error between the position stopped by the predicted facing control and the lamp center position, the facing setting position is a position laterally separated from the lamp center 60a. The front-facing camera 24 images the headlight L from an oblique direction. Therefore, a part of the lamp image 60 viewed from the original lamp center position as shown in FIG. 60 is lacking (in FIG. 59). This is because the lamp center 60a has not been correctly captured.

  If the Fresnel lens center 17a of the light receiving unit 11 is located at a position slightly deviated from the lamp center 60a when stopped by the predictive facing control, the image processing device 30 uses the lamp image information from the facing camera 24 to It is possible to determine the amount of deviation. Accordingly, in this state, assuming that the image processing device 30 processes the light distribution pattern information from the measurement camera 20 with the headlight tester 10 in the measurement mode, the image processing device 30 knows the amount of deviation. Corrections can be made as appropriate based on the amount. For example, as shown in FIGS. 58 and 59, when the shift amount is 10 mm on the right side, the light distribution pattern 33 imaged on the measurement screen 19 is also shifted 10 mm on the right side. Therefore, it is possible to eliminate the facing error by correcting the deviation amount of 10 mm on the right side. On the other hand, if there is a deviation amount of 10 mm to the right after the completion of the preliminary facing control, the light receiving unit 11 is again moved to the left by the deviation amount, and the position of the Fresnel lens center 17a of the light receiving unit 11 is set to be true. It is also possible to match the position of the lamp center 60b.

  Although specific embodiments of the present invention have been described in detail above, the present invention should not be limited to these specific embodiments, and various modifications can be made without departing from the technical scope of the present invention. Of course, it can be changed.

10: Headlight tester 11: Light receiving unit 17: Main lens (Fresnel lens)
19: Internal (for measurement) screen 20: Measurement camera 20b: Semiconductor color image sensor 21: Data processing device 24: Direct-facing camera 25: 10m Screen 30: Image processing device 33: Light distribution pattern (image)
37a: Horizontal cut line 37b: Diagonal cut line 38: Elbow point 40: Pixel 41: Zone 50: Reference sphere mounting device 51: Reference sphere 58: Illuminance meter

Claims (15)

An optical system that receives the projection light from the headlight,
A direct-facing imaging device that receives the projection light through the optical system and captures an image of the headlight;
An image processing apparatus for receiving the image data of the headlight from the imaging device for direct facing, binarizing means for converting the image data into binarized data, and another image around the center of the lamp of the headlight An image processing apparatus comprising: morphological processing means for performing morphological processing on the binarized data to form a shaped image separated from a portion; and center determining means for determining the center of the shaped image , A movement control means for performing movement control to face the headlight based on information on the center of the shaped image supplied from the image processing apparatus;
Headlight tester having.   The headlight tester according to claim 1, wherein the optical system has a main lens.   The headlight tester according to claim 2, wherein the main lens is a Fresnel lens.   The image processing apparatus is composed of a computer system including at least a CPU and a memory, and the binarization unit, the morphology processing unit, and the center determination unit control the operation of the headlight tester. The headlight tester according to any one of claims 1 to 3, wherein the headlight tester is stored in the image processing apparatus as a part of a program to be executed.   The headlight tester according to any one of claims 1 to 4, wherein the morphology processing means is processing means for performing a closing process a plurality of times until a shaped image having a desired size or shape is obtained.   The headlight tester according to any one of claims 1 to 5, wherein the center determining means determines the center of the shaped image as a directly-facing center. Furthermore, a screen on which the light distribution pattern of the projection light is projected through the optical system,
A measurement camera for imaging the light distribution pattern on the screen;
And the movement control means has a position of the optical axis of the headlight tester and the center of the shaped image when facing the headlight based on the information on the center of the shaped image. When the difference from the position falls within a predetermined allowable range, the movement control is stopped as being substantially opposed, and the image processing apparatus uses the difference in position as a correction amount from the measurement camera. The headlight tester according to claim 1, which is used when processing information on the light distribution pattern.   And a light source type determining means for determining a light source type of the headlight, wherein the measurement camera is a color camera and supplies information on the light distribution pattern to the image processing apparatus as RGB data. The light source type determining means determines the light source type of the headlight based on the RGB data, and the image processing device detects the headlight based on the light source type determination information and the lamp image information from the facing camera. The headlight tester according to claim 8, wherein the headlight tester identifies whether the projector type, the lens cut type, or the multi-reflector type. Receiving the projection light from the headlight and taking an image of the headlight,
Binarize the image,
Morphological processing is performed on the binarized image to form a shaped image separated from other image portions near the lamp center of the headlight,
Determine the center of the shaped image;
Moving the headlight tester to align the optical axis of the headlight tester with the center;
The headlight tester's confrontation method that encompasses that.   The method according to claim 9, wherein the image is captured using a high dynamic range (HDR) function of a facing camera.   11. The method of claim 10, wherein the image includes a darker image portion near the lamp center than its surrounding portion.   The method according to claim 9, wherein the morphology process is a closing process.   The method according to claim 12, wherein the closing process is repeated until the shaped image has a predetermined size or shape.   The headlight movement is stopped when the difference between the position of the optical axis and the position of the center of the shaped image falls within a predetermined allowable range. Method.   15. The method according to claim 14, further comprising applying the difference in position as a correction amount when imaging a light distribution pattern in which the projection light is projected on a screen and performing image processing based on information on the light distribution pattern. .