JP5590387B2 - Color target position determination device - Google Patents

Description

The present invention relates to a color target position determination apparatus that determines the position of a color target on a captured image by performing image processing on a captured image obtained by capturing a color target composed of a combination of different colors under a light source.

  In recent years, an increasing number of vehicles are equipped with a camera so that a driver of the vehicle can visually recognize a scene such as a side or rear of the vehicle via a monitor inside the vehicle. Furthermore, an apparatus that supports driving such as parking by performing image processing using a photographed image of the camera has been developed. In particular, with respect to a camera that captures a captured image serving as a base for creating information used for vehicle positioning, calibration such as optical axis adjustment is required with high accuracy. The target position specifying device is used for the calibration process of such an in-vehicle camera. For example, a marker (target) with a black and white checkered pattern placed in two places in the camera's field of view is photographed with an in-vehicle camera, the center point (calibration point) of the marker is detected through image processing, and the in-vehicle camera is A technique for calibrating is known (see, for example, Patent Document 1). With this technology, it is necessary to accurately detect the center point of the black and white checkered pattern, but depending on the illumination light and sunlight conditions, the edge detection accuracy of the black and white area becomes unstable, and the detection accuracy of the center point of the marker decreases. There is a problem of doing.

  In addition, in order to solve the problem that the detection ability of the specific color region becomes unstable depending on the state of the light source such as illumination light or sunlight, the light source is based on the color temperature estimated from the photographed image acquired by the in-vehicle camera. A conventional technique for reducing the influence is known (for example, see Patent Document 2). In other words, in order to estimate the color temperature of the light source outside the vehicle based on the captured image captured by the imaging unit mounted on the vehicle, an achromatic member is provided at a predetermined position in the vehicle. An image of the periphery of the vehicle is taken so as to include the sky. Estimating the color temperature of the achromatic member in the captured image, and the color temperature of one or more light sources existing in the sky in the captured image Is estimated. Then, based on each estimation result, the color temperature of the light source existing in the sky image area, which is within a predetermined color temperature determination range including the color temperature of the image area of the achromatic member, is determined as the final color temperature. It is assumed that In this technology, not only is it necessary to place an achromatic member at a predetermined position in the vehicle, but the achromatic member that becomes this color reference is contaminated or exposed to a local light source species. Therefore, it is impossible to accurately estimate the color temperature.

Japanese Patent Laying-Open No. 2008-131250 (paragraph numbers 0023-0040, FIG. 1) Japanese Patent Laid-Open No. 2009-271122, paragraph number [0009-0021], FIG. 1)

  In view of the above-described prior art, color effects generated in a photographed image obtained by photographing a color target composed of a combination of different colors under a light source can be used without requiring an additional achromatic member for color correction reference. There is a need for a technique that reduces and maintains color target position detection capabilities.

  A color target position determination apparatus according to the present invention includes an image input unit that inputs a captured image obtained by capturing a color target composed of a combination of different colors under a light source, and a light source estimation unit that estimates a light source type of the light source from the captured image. A color profile storage unit that stores a color profile created for each of the light source types so that a color target in the captured image has a characteristic that reduces the color influence of the light source for each of the light source types, and the light source estimation A color conversion unit that performs color correction on the captured image using a color profile that is selected and output from the color profile storage unit based on the light source type estimated by the unit, and a color-corrected captured image that is color-corrected by the color conversion unit The different color regions of the color target in And a color target location determination module for determining a position in.

  According to this configuration, the light source type received by the color target to be recognized from the captured image is first estimated from the input captured image, and the color influence such as color cast that the light source type has on the color target is reduced. A color profile having such characteristics is selected, and the color target in the photographed image input using the color profile is color-corrected. This color profile is created in advance so as to experimentally confirm the color influence on the color target for each light source type and reduce this color influence. Therefore, the captured image that has been color-corrected by the color conversion unit that uses an appropriate color profile that matches the estimated light source type is reduced in color influence to the extent that it is possible to identify different color areas of the color target. The subsequent color target position is determined smoothly. In particular, in a photographed image under a light source such as a sodium lamp, it is highly likely that the color other than yellow of the subject is determined to be gray or black, but by using a color profile that considers the color cast characteristics of the sodium lamp, A captured image capable of identifying different color regions of the color target is provided to the color target position determination module.

  As one of preferred embodiments when creating a color profile so as to reduce the color influence of the light source for each light source type, a plurality of primary colors (original color) constituting the color target Represents the original color under ideal sunlight, but it does not narrowly limit such color here, but has a predetermined color range that includes approximate colors of the original color. Each of the change colors (here, the change color) generated by moving the original color in the color space due to the color influence of the light source type (influence on the color) It is proposed to approximately obtain a common transformation matrix (having a predetermined color range including an approximate color of a specific change color) and to create the color profile based on an inverse matrix of this transformation matrix . For example, assuming that a color target has a pattern composed of a first color and a second color in a certain color space, the first color and the second color are moved in the color space by a special light source, and the third color and the second color are moved. If there are four colors, the first color changes to the third color (movement on the color space) and the second color changes to the fourth color (movement on the color space). be able to. Furthermore, when an inverse matrix of the transformation matrix is at least approximately obtained, a color profile can be created in advance so as to have a characteristic that reduces the color influence received by the light source based on the inverse matrix.

  If there are multiple types of color targets that must be recognized by the vehicle periphery image recognition device and the color configuration differs for each type, a dedicated color profile is created for each type of light source and for each type of color target. As a result, more accurate color correction is possible. Therefore, in one preferred embodiment of the present invention, the color conversion unit uses the color profile selected based on the estimated light source type and the characteristic color characterizing the color target to capture the captured image. It is configured to perform color correction.

It is explanatory drawing of the first half of the color correction principle which correct | amends the color influence by the light source employ | adopted with the color target position determination apparatus by this invention. It is explanatory drawing of the second half of the color correction principle which correct | amends the color influence by the light source employ | adopted with the color target position determination apparatus by this invention. It is explanatory drawing which shows an example of the arrangement | positioning relationship between the target used for camera calibration, and a vehicle. It is explanatory drawing which shows an example of a color target. Functional block diagram showing image processing function in color target position determination device It is a graph which shows the scanning result with respect to the target image in a YCbCr color system. It is explanatory drawing explaining the calculation method for calculating a threshold value. It is a flowchart which shows the flow of control in a color target position determination apparatus. It is a flowchart which shows a color target position determination process.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. First, the principle of the color correction technique that is used in the vehicle peripheral image recognition apparatus according to the present invention and corrects the color effect due to the light source will be described with reference to FIGS.
Here, it is assumed that the color target 2 is configured by a combination (pattern) of two different colors. As shown in FIG. 1A, one of the two colors (original colors including approximate colors) is named as the first color and expressed in C1 (R1, G1, B1) (RGB color system). Yes. The other of the two colors is named the second color and is represented by C2 (R2, G2, B2). That is, the characteristic colors that characterize the color target 2 are the first color and the second color, and accurate position detection is possible through detection of the boundary line between the first color and the second color. When the color target 2 is illuminated by a light source such as a tungsten lamp or a sodium lamp, the first color and the second color change to different colors (change colors including approximate colors). The color in which the first color is changed is named the third color and is represented by C3 (R3, G3, B3). The color in which the second color is changed is named the fourth color and is represented by C4 (R4, G4, B4). Each color change can be expressed as follows using a relational expression expressing the color change (transition) in a matrix.
C1 (R1, G1, B1) ・ M1 => C3 (R3, G3, B3) (1)
C2 (R2, G2, B2) ・ M2 => C4 (R4, G4, B4) (2)
A schematic diagram visualizing such a color change on the color system is shown in FIG. That is, the coordinate position of the original color is moved to the coordinate position of the change color by illumination with the light source, and the direction and distance of the movement are represented by conversion matrices: M1 and M2.
Next, it is considered that the color target 2 affected by the color influence of the light source is color-corrected so that it becomes the original color on the captured image. This correction can be realized by obtaining transformation matrices: M1 and M2 inverse matrices: M1 ⁻¹ and M2 ⁻¹ . In other words, the movement from the first color to the third color on the color system and the movement from the second color to the fourth color on the color system due to the color effect of the light source becomes the original color by reversing the movement in reverse. . This can be shown as follows.
C3 (R3, G3, B3) ・ M1 ^-1 => C1 (R1, G1, B1) (3)
C4 (R4, G4, B4) ・ M2 ^-1 => C2 (R2, G2, B2) (4)
A schematic diagram visualizing such color correction on the color system is shown in FIG.
In general, since it is difficult to obtain an accurate inverse matrix, it may be obtained approximately or experimentally or both. The color profile created so as to correspond to M1 ^-1 and M2 ^-1 obtained in this way has characteristics that at least reduce the color influence of the color target 2 on the specific light source.
In addition, a common transformation matrix is preferably created by optimally approximating a plurality of transformation matrices created individually for each color (first color and second color), and both colors (first and second colors) are obtained from the inverse matrix of the common transformation matrix. A color profile that matches one color and the second color may be created. Although substantially the same, a plurality of individually created inverse matrices are preferably optimally fused into one inverse matrix, and both colors (first color and second color are selected from the fused inverse matrix. ) May be created.

When a plurality of light sources that have a color effect such as strong color cast have to be considered, it is necessary to create a color profile as described above for each light source and store it in a selectable manner. Furthermore, it is necessary to perform light source estimation as to what kind of light source is illuminated from the captured image to be processed. For this light source estimation, it is advantageous to obtain a captured image under the condition that each light source illuminates the color target 2 and derive it from the image feature amount of the captured image. In order to realize such a method, a knowledge base or neural network technique that can estimate the light source type from the image characteristics through an experimental learning process is preferable, but other well-known algorithms are used. May be. As schematically shown in FIG. 2, an image feature amount of a captured image to be processed is input, an estimated light source type of a light source is calculated and output, and a color profile (inverse matrix) adapted to the light source type is a color profile storage unit. The color target 2 is obtained by making it easy to identify and recognize the constituent colors by correcting the color of the captured image using the output color profile.
In addition, when it is necessary to process a plurality of types of color targets 2 having different color configurations, a color profile (inverse matrix) for each of the plurality of types of color targets 2 may be created.

  Next, an embodiment of the color target position determination apparatus of the present invention, which employs the principle of the color correction technique for correcting the color influence caused by the light source, will be described with reference to the drawings. FIG. 3 is an explanatory diagram illustrating an example of an arrangement relationship between the calibration target 2 (calibration index) and the vehicle 1. The camera 11 (on-vehicle camera) is installed above the license plate behind the vehicle 1 at a position offset laterally from the longitudinal axis of the vehicle body with the camera optical axis facing downward (for example, 30 ° downward from the horizontal). . In the illustrated example, the longitudinal axis of the vehicle body and the camera optical axis are not parallel. The camera 11 is, for example, a wide-angle camera with a horizontal viewing angle of 110 to 120 °, and can capture a region up to about 8 m behind. The camera 11 is calibrated to absorb mounting errors when mounted on the vehicle 1 at a vehicle production factory or the like. Further, in a repair shop or the like, calibration is performed so as to correct the displacement of the camera 11 due to accumulation of vibration caused by running or impact. Examples of attachment errors to be absorbed include roll angle, pan angle, tilt angle, and attachment position. Hereinafter, a description will be given of an example of a camera attachment error determination operation for camera calibration in an automobile production factory.

  As shown in FIG. 3, the calibration of the camera 11 is performed in a state where the vehicle 1 is stopped at a predetermined position. For example, if the vehicle 1 is moved back or forward and the wheels are stopped by a tire groove or a tire stop provided at a predetermined position, the vehicle 1 can be stopped at an accurate position. In the example shown in FIG. 3, two color targets (hereinafter simply referred to as targets) 2 (2a, 2b) are arranged on the floor surface. As is apparent from the figure, the distance between the two targets 2 a and 2 b is narrower than the tread width of the vehicle 1, and the target 2 is not easily stepped on by the wheels of the vehicle 1. In contrast to this arrangement, an arrangement may be adopted in which the distance between the two targets 2a and 2b is wider than the tread width of the vehicle 1 so that the target 2 is not easily stepped on by the wheels of the vehicle 1.

  In FIG. 3, the vehicle 1 is stopped so that the floor surface at the center of the rear end is the origin Ow of the world coordinate system (reference coordinate system, Xw, Yw, Zw). Here, the axes of the camera coordinate system (Xc, Yc, Zc) centered on the optical center Oc of the camera 11 are not parallel to the world coordinate system. Both the world coordinate system and the camera coordinate system are right-handed coordinate systems, and in the drawing, the Xw axis in the vertical direction and the Xc axis in the substantially vertical direction are not shown in the drawing. Coordinate conversion between the world coordinate system and the camera coordinate system can be performed using a well-known calculation method.

  The targets 2 are arranged at least at two locations within the visual field range of the camera 11. The target 2 is arranged so that its coordinates are known in the world coordinate system. In this example, as shown in FIG. 4, the target 2 has a blue and red checkered pattern as characteristic colors that characterize the target 2. A point Q at the center of the pattern is a calibration point, which is a reference for calibration of the camera 11. That is, the target 2 is arranged so that the coordinates of the calibration point Q are known in the world coordinate system. In addition, although the example of a total of four rectangles of two blue rectangles and two red rectangles was shown here, there may be a total of four or more, and the number and shape are shown here. It is not limited to.

  The target 2 is disposed directly or indirectly on a floor surface exposed to ambient light mainly composed of a lamp light source such as a fluorescent lamp or an incandescent lamp. It becomes difficult. Therefore, in the present invention, by adopting a coating method in which a matte satin-like paint is applied twice to the iron plate, the surface of the target 2 is made to be in a matte state, and a photographed image that is not easily affected by reflection of ambient light or the like can be obtained. I have to.

  In the example shown in FIG. 3, the two targets 2 are arranged on the floor symmetrically with respect to the main axis of the vehicle (Zw axis of the world coordinate system) (D1 = D2, W1 = W2). However, it is not always necessary to be left-right symmetric, and if the coordinate value is known within the field of view of the camera 11, its arrangement is free. That is, it is possible to arbitrarily set the arrangement of the target 2 depending on the area that can be secured in the factory and the relationship with other equipment.

  The dimension of the target 2 is appropriately determined so that the calibration point Q can be detected with high accuracy according to the resolution of the camera 11, the performance of the image processing function for processing the image taken by the camera 11, the arrangement position of the marker, and the like. As an example, when D1 and D2 are about 1 to 2 m and W1 and W2 are about 0.5 m, a target 2 having 10 to 15 cm square for each black and white and 20 to 30 cm square as a whole is shown in FIG. Used.

  In this embodiment, the target position specifying apparatus according to the present invention is substantially constituted by an image processing unit having a computer as a core member. FIG. 4 is a block diagram schematically showing the image processing function. Has been. As shown in FIG. 4, the image processing unit includes an image input unit 31, a color conversion module 4, a target position determination module 5, and an attachment accuracy determination module 32. The image input unit 31 inputs a captured image including the target 2 acquired by the camera 11 and develops it in a working memory (not shown). The target position determination module 5 obtains the position of the target 2, particularly the position of the calibration point Q of the target 2 from the captured image developed in the working memory. The attachment accuracy determination module 32 determines the attachment accuracy of the camera 11 from the difference between the position of the calibration point Q of the target 2 specified by the target position determination module 5 and the target calibration point position.

The color conversion module 4 includes a light source estimation unit 41, a color profile storage unit 42, a color profile selection unit 43, and a color conversion unit 44. The light source estimation unit 41 inputs various image feature amounts calculated from the captured image input through the image input unit 31, for example, average luminance, histogram characteristics, color component ratios, etc., thereby calculating weights, rules, etc. The estimated light source type is output through the operation. The weight calculation here is, for example, a calculation that assigns a weight coefficient to each input parameter and corrects the weight coefficient by iterative learning to minimize the error in the output result, as used in neural networks, for example. It is a general term for operations using expressions. A rule operation is a general term for operations that derive a result based on a predetermined rule such as an ifthen statement. The content of the rule itself is, for example, “If the B (blue) color component value is higher than a predetermined value and the R (red) and G (green) component values are lower than the predetermined value, the probability of an orange lamp is predetermined%. And so on. "
In addition, although a tungsten lamp, a sodium lamp, a fluorescent lamp, an LED lamp, etc. are mentioned as a light source seed | species, it is good also as light source seed | species estimated here also including in sunlight. Also, the same light source type may be classified according to the intensity of each lamp and the magnitude of color influence.

The color profile storage unit 42 stores a color profile (color matrix) used as a correction table for reducing the color influence such as color fog from the image of the target 2 as much as possible, which is used for the color correction described with reference to FIG. ing. In this embodiment, each color profile is created for each combination of a specific light source type and a specific color configuration of the target 2 (here, blue and red). It is made into a database so that it can be searched and extracted as a search keyword. Further, for this target 2, a color profile (blue correction inverse matrix) for reducing the color influence on blue by a specific light source type and a color profile (red correction inverse matrix) for reducing the color influence on red are used. A combined blue / red common color profile (common inverse matrix) is used. This blue / red common color profile is obtained by approximating a common transformation matrix that fuses a transformation matrix that reduces the color effect on blue and a transformation matrix that reduces the color effect on red, and the inverse of the common transformation matrix. It is also possible to create based on a matrix.
The color profile selection unit 43 selects a suitable color profile from the light source type estimated by the light source estimation unit 41 and the preset color configuration of the target 2, here, the blue and red checkered pattern color configuration. To the color conversion unit 44. The color conversion unit 44 performs color correction of the captured image using the color profile selected by the color profile selection unit 43.

  In this embodiment, the target position determination module 5 includes a preprocessing unit 51, a color difference conversion unit 52, a threshold setting unit 53, a color area determination unit 54, a boundary detection unit 55, and a target position calculation unit 56. Contains. The pre-processing unit 51 performs image distortion correction and level adjustment (density correction, contrast correction, etc.) caused by the lens characteristics of the camera 11 as necessary.

The color difference conversion unit 52 calculates the RGB pixel value of the captured image color-corrected by the color conversion unit 44 to obtain a blue component value, a red component value, and a luminance value, and generates color difference image data. For this conversion, for example, the following expression used in the conversion from the RGB color system to the YCbCr color system can be used.
Y = 0.299 x R + 0.587 x G + 0.114 x B
Cb = 0.564 × (B−Y)
Cr = 0.713 × (R−Y)
Here, Y: luminance value, Cb: blue component value, Cr: red component value, R: R pixel value, G: G pixel value, B: B pixel value. As can be understood from this equation, the first color component value and the second color component value, for example, the blue component value: Cb and the red component value: Cr change according to the luminance (brightness) of the photographed image. It is.

  The threshold setting unit 53 determines whether the target pixel (target region) is red or blue based on the luminance value (Y) included in the color difference image data generated by the color difference conversion unit 52. A specific color detection threshold is set as a determination condition. This is because the luminance value affects the blue component value and the red component value in the YCbCr color system, and the component value increases as the captured image becomes brighter, and the component value decreases as the image becomes darker. FIG. 6 shows how the blue component value: Cb and the red component value: Cr change when the color difference image including the target 2 is scanned in the horizontal direction. FIG. 6A is based on a photographed image acquired in a bright environment. The luminance value: Y is relatively high, and the levels of the blue component value: Cb and the red component value: Cr are also relative. It is getting higher. FIG. 6B is based on a photographed image acquired in a dark environment, and its luminance value: Y is relatively low and its luminance value: Y is relatively low. Therefore, in order to accurately detect the blue region or red region of the target 2 from the color difference image data, when the luminance value is high as shown in FIG. 6A, thresholds (THb, THr as detection conditions) ) Level should be relatively high. Conversely, when the luminance value is low as shown in FIG. 6A, the threshold level as the detection condition may be relatively lowered.

As one of the preferred methods for dynamically changing the threshold level according to the brightness (luminance level) of the image as described above, as schematically shown in FIG. Average value of color component values to be determined: VA_avg and pixels whose color component values to be determined belong to a higher level among all pixels in the captured image (for example, have a color component value of 95% or more of the maximum color component value) From the average value VH_avg of the upper level corresponding to the color component corresponding red area of the color component value of pixel) and the difference value thereof: ΔV_avg (= VH_avg−VA_avg), the pixel of the photographed image is in a specific color area (here, The specific color threshold value TH (THb, THr) for determining that it belongs to either of (blue or red) can be obtained by the following equation.
TH = VA_avg + K × ΔV_avg
Here, K is a constant value and is a value obtained experimentally in advance.
The threshold value TH calculated in this way is derived by a function including the luminance component in the captured image as a parameter of Cb: blue component value or Cr: red component value, and therefore the luminance level in the captured image is high. And increases when the luminance level is low. In order to simplify the calculation processing for obtaining this threshold value, it is preferable to obtain a common threshold value by obtaining one of the blue threshold value: THb or the red threshold value: THr and using this as the other threshold value.

Various methods other than those described above can be considered as a method of dynamically changing the threshold level according to the brightness (luminance level) of the photographed image. The upper concept can be explained as follows. it can. For example, the threshold value THb for determining the blue region is Y: statistical (average value, median value, etc.) numerical value: f (Y) and Cb: blue component value. It can be obtained by the following function: Fb with a parameter of statistical numerical value: f (Cb).
THb = Fb (f (Y), f (Cb))
The threshold value for determining the red region: THr is Y: statistical value of the luminance value of the photographed image or the luminance value of the color difference information: f (Y) and Cb: statistical value of the red component value: f (Cr) It is possible to obtain it with the following function: Fr.
THr = Fr (f (Y), f (Cr))
Here, the two threshold values of the threshold value for determining the blue region: THb and the threshold value for determining the red region: THr are obtained, but it is also possible to use one threshold value in combination.

  The color area determination unit 54 sequentially scans the color difference image including the target 2 using the threshold value set dynamically by the threshold setting unit 53, that is, for each captured image, and determines a blue area and a red area. . At this time, as a determination condition for determining the blue region, it is preferable to set the AND condition that the blue component value exceeds the threshold and the red component value falls below the threshold. Similarly, as a determination condition for determining the red region, an AND condition may be that the red component value exceeds the threshold and the blue component value falls below the threshold. That is, in the color difference signal, since the blue component value and the red component value are in a counter electrode relationship, more accurate blue region determination can be performed by using the difference. However, in the present invention, when determining the blue region and the red region, it is limited to the AND condition that one color component value exceeds the threshold value and the other color component value falls below the threshold value. I'm not. As a determination condition for determining the blue region, only that the blue component value exceeds the threshold value may be employed, or as a determination condition for determining the red region, only that the red component value exceeds the threshold value is employed. You may do it.

  The boundary detection unit 55 detects the boundary between the blue region and the red region of the target using the determination result of the blue region and the red region by the color region determination unit 54. Since the boundary detected by the boundary detection unit 55, that is, the intersection of two boundary lines becomes the calibration point Q, the target position calculation unit 56 determines the position of the target 2 in the captured image based on the boundary detection result by the boundary detection unit 55. That is, the calibration point Q can be calculated.

  A captured image acquired by the camera 11 and developed in the frame memory, and an image indicating a deviation between the calibration point Q specified by the target position determination module 4 and the target calibration point position are displayed on the monitor 12 through the image signal generation unit 33. It is also possible to do. At this time, since a human can understand the color cast caused by the light source while correcting the color cast to a certain degree of color, a mode is also provided in which the monitor 12 displays a photographed image that has not been color corrected through the color conversion unit 44. It is convenient.

Next, the flow of control in the color target position determination apparatus including the color conversion module 4 and the target position determination module 5 configured as described above will be described with reference to FIGS. FIG. 8 shows color conversion control, and FIG. 9 shows target position determination control.
First, the vehicle 1 is accurately positioned and stopped at a predetermined position of the inspection site (# 01). After confirming that the vehicle has stopped, the camera 11 is operated to photograph the vicinity of the vehicle (# 02). The camera is set so that the image taken by the camera 11 of the vehicle 1 stopped at a predetermined position includes the images of the two targets 2 even if the mounting accuracy of the camera 11 is somewhat poor. The captured image (RGB color image data) input through the image input unit 31 is subjected to basic image processing such as correction of image distortion caused by lens characteristics (# 03).

  Next, in order to estimate under what light source the captured image was captured, a plurality of image feature quantities for light source estimation are calculated from the captured image (# 04). The light source type estimation calculation is performed using the calculated image feature quantity as an input value, and the light source type is output (# 05). The color configuration of the target 2 to be processed is read (# 06). A matching color profile is selected using the output light source type and the read color configuration as a search keyword (# 07). Using the selected color profile, color correction is performed on the captured image (# 08). The color-corrected captured image is developed in the working memory and is subjected to target position determination processing by the target position determination module 5 shown in FIG. 9 (# 10).

In the target position determination process, first, conversion from RGB color image data to YCbCr color (color difference) image data is performed (# 11).
Next, the threshold setting unit 53 converts the Cb: blue component value or the Cr: red component value of the YCbCr color image data to the above-described equation,
TH = VA_avg + K × ΔV_avg
And setting a detection threshold for the blue region of the target 2 (# 12) and a detection threshold for the red region of the target 2 (# 13). According to the acquired captured image, the YCbCr-based captured image is sequentially scanned using the threshold value set through the arithmetic processing in step # 08, and the blue region determination (# 14) of the target 2 and the red region are detected. Determination (# 15) is performed. In the determination of the blue region, the determination condition is that the blue component value exceeds the blue detection threshold and the red component value falls below the red detection threshold. Similarly, in the determination of the red region, the determination condition is that the red component value exceeds the threshold value and the blue component value falls below the threshold value.

  Next, a boundary line between the blue region and the red region of the target 2 is detected from the determination result of the blue region and the red region (# 16). Of course, it is also possible to simultaneously determine the blue region and the red region of the target 2 and detect the boundary line between the blue region and the red region. In any case, since the boundary line between the detected blue region and red region shows a form in which almost two lines are orthogonal, the calibration point coordinates are calculated with the intersection as a calibration point (# 17). .

  Through the above processing steps, the position of the target 2, that is, the coordinate position of the calibration point can be obtained. Therefore, the deviation amount between the preset target calibration point and the calibration point calculated in step # 17 is calculated (# 18). Based on the calculated deviation amount, the mounting accuracy determination module 32 determines the mounting accuracy of the camera 11 (# 19).

[Another embodiment]
(1) Each functional unit in the color conversion module 4 and the target position determination module 5 described above indicates a sharing as a function, and does not necessarily have to be provided independently. Of course, each function may be realized by cooperation of hardware such as a microcomputer and software such as a program executed on the hardware.
(2) In the embodiment described above, the target 2 to be processed by the color target position determination device is the target 2 for determining the in-vehicle camera mounting position, but the target for stopping at a parking lot or a battery charging station. It may be. Also, the present invention may be applied by regarding the white line or yellow line drawn on the road as the target 2.
(3) In the above embodiment, the color profile for reducing the color influence on the first color (blue) and the second color (red) as the color profile having the characteristic of reducing the color influence on the target 2 by the light source. A common blue / red color profile was used, which was combined with a color profile to reduce the color effect on the image. Instead, a blue correction color profile for reducing the color influence on blue and a red correction color profile for reducing the color influence on red may be prepared, and these may be applied individually.

  The present invention can be widely used for an image processing technique for reducing the color influence of a light source exerted on a subject characterized by a combination of different colors.

2: Color target 31: Image input unit 32: Mounting accuracy determination module 4: Color conversion module 41: Light source estimation unit 42: Color profile storage unit 43: Color profile selection unit 44: Color conversion unit 5: Target position determination module

Claims (3)

An image input unit for inputting a photographed image obtained by photographing a color target composed of a combination of different colors under a light source;
A light source estimation unit for estimating a light source type of the light source from the captured image;
A color profile storage unit that stores a color profile created for each of the light source types so that the color target in the captured image has a characteristic that reduces the color influence of the light source for each of the light source types;
A color conversion unit that color-corrects the captured image using a color profile selected and output from the color profile storage unit based on the light source type estimated by the light source estimation unit;
A color target location determination module for determining the position in the captured image of the color target the different color regions of the color target in the color-corrected color correction captured image by the color conversion unit identification determined by,
Equipped with a,
From the plurality of primary colors constituting the color target, a common conversion matrix for each changed color generated by moving the primary color in the color space due to the color influence of the light source type is approximately obtained, and this conversion is performed. A color target position determination apparatus in which the color profile is created based on an inverse matrix of a matrix . An image input unit for inputting a photographed image obtained by photographing a color target composed of a combination of different colors under a light source;
A light source estimation unit for estimating a light source type of the light source from the captured image;
A color profile storage unit that stores a color profile created for each of the light source types so that the color target in the captured image has a characteristic that reduces the color influence of the light source for each of the light source types;
A color conversion unit that color-corrects the captured image using a color profile selected and output from the color profile storage unit based on the light source type estimated by the light source estimation unit;
A color target position determination module that identifies and determines different color regions of the color target in the color-corrected captured image that has been color-corrected by the color conversion unit, and determines the position of the color target in the captured image;
With
The color conversion unit is a color target position determination device that performs color correction on the captured image using a color profile selected based on the estimated light source type and a characteristic color that characterizes the color target. The color target position determining apparatus according to claim 1 , wherein the color conversion unit performs color correction on the captured image using a color profile selected based on the estimated light source type and a characteristic color characterizing the color target. .

Images