JP5601913B2 - Vehicle detection device - Google Patents

Description

  The present invention relates to a vehicle detection device that images a surrounding area of a host vehicle and detects another vehicle based on the captured image.

  In order to improve the visibility of the front area of the vehicle, it is preferable that the headlight of a vehicle such as an automobile emits light with a so-called high beam light distribution, but in this high beam light distribution, it exists in the front area of the own vehicle. There is a risk of dazzling preceding and oncoming vehicles. For this reason, in Patent Document 1, it is detected whether or not an irradiation prohibited object such as a preceding vehicle or an oncoming vehicle exists in the front area of the vehicle, and irradiation with high beam light distribution is performed on the area where the irradiation prohibited object exists. Forbidden, and other areas are irradiated with high beam light distribution to ensure the visibility of the front area of the host vehicle while preventing the dazzling of the preceding and oncoming vehicles. Further, in Patent Document 2, the front area of the host vehicle is imaged with a camera, the vehicle position of another vehicle existing in the front area is detected from the obtained image, and the detected vehicle position is irradiated with low beam light distribution. Further, by irradiating the undetected vehicle position with the high beam light distribution, the visibility of the front area of the own vehicle is ensured while the dazzling with respect to other vehicles is prevented.

JP 2008-37240 A JP 2010-957 A

When other vehicles existing in the front area of the own vehicle are detected when controlling the light distribution of the headlamps as in Patent Documents 1 and 2, the front area of the own vehicle is detected with a camera as in Patent Document 2. A technique is used in which another vehicle is detected by taking an image and analyzing the obtained image. At this time, it is determined whether the light spot reflected in the captured image is light from a vehicle illumination device such as a preceding vehicle or an oncoming vehicle or light from a fixed illumination device such as a building or a fixed sign lamp. For this purpose, for example, a light spot in an image is detected, and the light spot is detected as an attribute such as the size, shape, color, distribution, movement locus, etc. It is determined whether it is due to the light from the tail lamp or the light from the fixed illumination device. However, in this method, since it is necessary to make a determination for all the light spots captured in the captured image, the number of data to be processed is large, and the number of man-hours for the determination is extremely large. Therefore, it becomes difficult to quickly detect the preceding vehicle and the oncoming vehicle, and it becomes difficult to control the light distribution of the headlight of the host vehicle in real time.

  An object of the present invention is to provide a vehicle detection device capable of detecting a light spot reflected in a captured image and detecting another vehicle quickly and with high accuracy based on the size of the detected light spot. It is.

The present invention is a vehicle detection device provided with other vehicle detection means for detecting a light spot in an image captured by the own vehicle and detecting another vehicle based on the light spot, and focuses a predetermined region of the image. In-focus area setting means for setting the other areas as non-focus areas, and the other vehicle detection means detects the size of a clear light spot in the focus area and a blurred light spot in the non-focus area. The light spot is detected as another vehicle when the detected size is smaller than the reference value. The focusing area setting unit may change the focusing area according to the vehicle speed and the steering direction of the host vehicle.

Further, the focusing area setting unit sets the focusing area so as to include the other vehicle closest to the host vehicle among the detected other vehicles. Alternatively, the focusing area setting unit may set the focusing area at the farthest forward position of the own vehicle when there is no other vehicle detected.

  The vehicle detection device of the present invention images the tail lamps and headlamps of other vehicles in the in-focus area by setting the area where the other vehicle exists as the in-focus area and setting the other area as the out-of-focus area. The light spot obtained in this way is obtained as a clear image, and the light spot obtained by imaging illumination light from a fixed illumination device or the like in the out-of-focus area is obtained as an image enlarged by defocusing. Therefore, by detecting the light spot of the image and comparing the size, it is possible to clearly determine whether the light spot is that of another vehicle or that of a fixed lighting device. Vehicle detection can be realized quickly and with high accuracy compared to the case of detection.

The conceptual block diagram of the motor vehicle provided with the vehicle detection apparatus of embodiment of this invention. The block block diagram of a focusing area | region setting part. The schematic diagram of a focusing pattern. The schematic diagram of the light spot detected with the image in which the focusing area | region was set. The flowchart which shows the operation | movement of other vehicle detection.

  Next, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a conceptual configuration diagram of an embodiment in which a vehicle detection device of the present invention is applied to a headlight control device that controls light distribution of a headlamp. In the car CAR, an imaging camera CAM for imaging the front area of the host vehicle, and a vehicle detection that detects other vehicles by detecting a light spot captured in an image obtained by imaging with the imaging camera CAM. A device 1 is provided. The headlamp control device 2 controls the light distribution of a headlamp (headlamp) HL that illuminates the front area of the host vehicle corresponding to the other vehicle detected by the vehicle detection device 1. It is configured as. Further, the vehicle detection device 1 is inputted with respective sensor outputs of a vehicle speed sensor Sv for detecting the vehicle speed of the host vehicle and a steering sensor Sw for detecting the steering direction of the host vehicle.

Here, the headlamp HL can be switched between high beam distribution and low beam distribution under the control of the light distribution controller 2. In order to switch the light distribution, the headlamp HL includes, for example, a shade (light-shielding plate) for setting the low beam light distribution, which is not shown, and as a lamp for setting the high beam light distribution by driving the shade. Although this type of headlamp has already been proposed, the description thereof will be omitted here. Alternatively, a headlamp in which a plurality of lamp units having different light distributions are combined and the light distribution is switched by selecting and lighting these lamp units may be used.

The imaging camera CAM can use a digital camera equipped with an imaging device that has been used so far. Here, a CCD imaging device or a MOS imaging device that outputs an image signal corresponding to the captured image is used. It consists of a digital camera.

The vehicle detection device 1 performs signal processing on an image captured by the imaging camera CAM and sets a focus area, and a light captured by image analysis of the captured image. The light spot detection unit 12 for detecting a point, and the detected light spot attribute, here, the color and size of the light spot (the diameter and vertical and horizontal dimensions of the light spot) are detected, and further detected based on the detected color Another vehicle detection unit 13 is provided that detects other vehicles by comparing the size with a reference value.

  Here, as shown in FIG. 2, the focus area setting unit 11 includes a focus pattern storage unit 111 that stores a preset focus pattern, and a plurality of focus pattern storage units 111 that store the focus pattern storage unit 111. A focusing processing unit 112 that sets only a part of the image captured based on the focusing pattern to a focused state and sets other areas to a non-focused state (out-of-focus state). The plurality of in-focus patterns in the in-focus pattern storage unit 111 are, for example, illustrated as dotted in FIGS. 3A to 3C, for example, a straight road pattern P1, an expressway pattern P2, and left and right curved road patterns ( Here, only the left curve road pattern P3 is illustrated), and an area corresponding to the captured image in advance based on the angle of view of the imaging camera CAM mounted on the host vehicle, the camera optical axis position, the camera mounting position, and the like. The pixel position information of the in-focus area in the image corresponding to a straight road, a highway, a curved road, etc. in the captured image is output.

Further, the focusing processing unit 112 regards an image input from the imaging camera CAM as an in-focus signal for an image signal in a focusing area set by the focusing pattern output from the focusing pattern storage unit 111. The image signal outside the in-focus area is output as a signal that has undergone defocus processing. This focusing processing unit 112 outputs a moving average calculating unit 113 that calculates a moving average for the image signal, and outputs the image signal as it is based on the focusing pattern, or whether to input to the moving average processing unit 113. And a selection unit 114 for selection. That is, the selection unit 114 is controlled based on the pixel position information of the focus pattern, and the image signal in the focus area is output as it is, but the image signal outside the focus area is input to the moving average calculation unit 113. To do. As a result, a clear image signal as captured is obtained within the in-focus area, but the level difference between the high-intensity part and the low-intensity part is reduced by moving and averaging the image signal outside the in-focus area. As a result, the signal of a blurred image with reduced sharpness is obtained. Alternatively, the technique described in Japanese Patent Application Laid-Open Nos. 4-81176 and 11-41512 can be applied as a technique for obtaining the same blur effect.

The light spot detection unit 12 performs image analysis on an image in which the in-focus area is set by the in-focus area setting unit 11, that is, an in-focus area, and an out-of-focus area out-of-focus image. The reflected light spot is detected. For example, the image output from the focus area setting unit 11 is scanned, and a light spot is detected from a pixel having relatively high luminance obtained from the scanning signal. At this time, a clear light spot in the in-focus state is detected in the in-focus area, but outside the in-focus area, the light spot in the out-of-focus state, that is, the blur increases the diameter size and the vertical and horizontal dimensions compared to the in-focus state. A light spot is detected.

The other vehicle detection unit 13 detects the color and size of the light spot detected by the light spot detection unit 12. The colors are red corresponding to the tail lamp of the preceding vehicle and white and yellow corresponding to the head lamp of the oncoming vehicle. For the size, the diameter and vertical / horizontal dimensions of the light spot are detected, and the detected size is compared with a reference value. This reference value is based on the standard size of the light spot when the tail lamp of the preceding vehicle and the headlamp of the oncoming vehicle are photographed in focus, preferably the size of the light spot when the host vehicle is closest to another vehicle. When the detected light spot is smaller than this reference value, it is detected that the vehicle is another vehicle.

  The operation of the vehicle detection apparatus having the above configuration will be described with reference to the schematic diagram of FIG. 4 and the flowchart of FIG. 5 in addition to FIG. The focusing area setting unit 11 inputs the vehicle speed and steering angle of the host vehicle (S11), and selects a focusing pattern corresponding to the traveling state of the host vehicle from the focusing pattern storage unit 111 (S12). Here, one of the focusing patterns P1 to P3 shown in FIG. 3 is selected and input, and a focusing area is set corresponding to the input focusing pattern. That is, the straight road pattern P1 set as a default during normal driving is input, and the area corresponding to the straight road in the image captured as shown in FIG. Is the out-of-focus area.

Next, the front area of the host vehicle is imaged by the imaging camera CAM, and the imaged image is input to the focusing area setting unit 11 (S13). The image signal in the straight road pattern P1 is directly output by the selection unit 114 based on the focus pattern selected at this time, that is, the straight road pattern P1 (S14), and the image signal outside the straight road pattern P1 is a moving average. The image signal is input to the calculation unit 113 and output as a moving average (S15). In the image obtained in this manner, the light spot LD is imaged in the focused area A1, that is, the road corresponding area corresponding to the straight road pattern P1, and the light spot LD is captured outside the focused area A1, that is, the road corresponding area. The point LD is in a state of being imaged in a defocused state.

  In this way, the light spot detection unit 12 detects all the light spots LD in the image for which the focus area setting processing has been performed (S16). As shown in FIG. 4A, the light spot LD is detected as a clear light spot in the in-focus area A1 surrounded by the broken line, so that the light spot LD is detected outside the focus area A1. Since the image is taken out of focus, it is detected as a light spot blurred due to blur. And the size of the light spot LD detected in the other vehicle detection part 13 is compared with a reference value (S17). In other words, the light spot LDi existing in the in-focus area A1 is focused, so the size is smaller than the reference value, and the light spot LDo existing outside the focus area A1 is smaller than the reference value due to out-of-focus. Therefore, the light spot of the other vehicle or the light spot of the fixed illumination device can be determined based on the size, and the light spot having a size smaller than the reference value is detected as the other vehicle (S18). Otherwise, it is detected as not being a vehicle (S19). Furthermore, although illustration in a flowchart is abbreviate | omitted, the other vehicle detection part 13 detects the color of the detected light spot LDi. When the light spot LDi is red, it is a tail lamp. When the light spot LDi is white, it is a head lamp. From this, it is possible to detect a preceding vehicle and an oncoming vehicle. Therefore, as compared with the case where each attribute is detected for all the light spots as in the prior art and other vehicles are detected by performing various comparisons on the detected attributes, the other vehicles are detected only by the color and size of the light spots. It is possible to detect the preceding vehicle and the oncoming vehicle with high accuracy.

  When the vehicle speed from the vehicle speed sensor Sv changes and the host vehicle is traveling at a predetermined speed or higher, that is, when traveling at high speed, the focus pattern storage unit 111 changes the focus processing unit 112 to FIG. The indicated highway pattern P2 is input. As a result, as shown by a broken line in FIG. 4B, the focusing area A2 is changed to an area corresponding to the far front area of the own vehicle, and the tail lamps and heads of the preceding vehicle and the oncoming vehicle existing far ahead of the own vehicle. The light spot detection unit 12 detects the light spot LDi of the lamp in a focused state. The other vehicle detection unit 13 detects the other vehicle by changing the reference value of the detected light spot to a small value and comparing the size of the detected light spot LDi with this reduced reference value. Therefore, even if there are light spots LDo generated by a relatively small sized fixed illumination device such as a delineator existing far in front of the host vehicle, these light spots LDo are increased in size due to defocusing. The light spot LDo can also be discriminated, and other vehicles existing far ahead can be detected with high accuracy.

  During normal traveling or high-speed traveling, when the steering signal from the steering sensor Sw changes and the vehicle is turning, that is, during curve traveling, the focusing pattern storage unit 111 to the focusing processing unit 112 The curve road pattern P3 (here, the left curve road pattern) corresponding to the steering direction shown in FIG. As a result, as shown by a broken line in FIG. 4C, the in-focus area A3 is changed to an in-focus pattern in which an area bent to the left is added to the straight forward area of the own vehicle. The light spot LDi of the preceding and oncoming cars traveling on a curved road ahead is detected in focus, and the light spot LDo of other fixed lighting devices is detected out of focus. The other vehicle detection unit 13 can detect the other vehicle by comparing the size of the light spot LD with the reference value.

In this embodiment, information on other vehicles such as a preceding vehicle and an oncoming vehicle detected by the other vehicle detection unit 13 is input to the headlight control device (light distribution control unit) 2. The light distribution of the headlamp HL is controlled corresponding to the information. For example, when the other vehicle is detected, the headlamp HL is set to the low beam light distribution, and when the other vehicle is not detected, the high beam light distribution is set. Thereby, while preventing the dazzling of other vehicles, the front area | region of the own vehicle is illuminated brightly, visibility is improved, and safe driving | running | working is ensured.

  As described above, in the vehicle detection device 1 of the embodiment, by setting the focus region for the image captured by the focus region setting unit 11, the light spot of the other vehicle in the image becomes a small size, and the fixed illumination device The light spot becomes a large size, and these light spots are detected by the light spot detection unit 12, and the other vehicle detection unit 13 determines the size of these light spots, thereby detecting the other vehicle only by the size of the light spot. Is possible. Therefore, it is not necessary to perform the process of detecting the other vehicle by performing discrimination for various attributes of the light spot detected from the captured image as in the conventional case, and the other vehicle can be detected quickly. This also makes it possible to quickly perform light distribution control of a suitable headlamp corresponding to another vehicle.

  In the embodiment, the focus area setting unit 11 sets the focus area based on the focus pattern stored in the focus pattern storage unit 111, but the information on other vehicles detected in the previous process is used. You may make it set the focusing area | region in the next process. That is, at first, as described above, the in-focus area is set by using the in-focus pattern and the other vehicle is detected. After the other vehicle is detected once, the other vehicle that is closest to the host vehicle is detected. An existing area is set as a focusing area. As a result, a detection state equivalent to the case shown in FIG. 4A is obtained, and thereafter, the other vehicle can be reliably tracked and detected, and optimal lighting control for the other vehicle and the host vehicle is possible. Become. Further, when no other vehicle is detected in the previous process, the in-focus area is set as an area corresponding to the farthest position of the own vehicle. As a result, a detection state equivalent to the case shown in FIG. 4B is obtained, and it becomes possible to quickly detect other vehicles newly appearing in front of the host vehicle.

  The focus area setting unit 11 according to the embodiment sets and changes the focus area by performing signal processing on the captured image, but can be performed optically. For example, a soft focus lens can be attached as an imaging lens of the imaging camera CAM. When the lens is replaced with a soft focus lens, a region near the lens optical axis is set as a focusing region, and the surrounding region is set in a soft focus state, that is, in focus. You may make it be in a blurred state. When this is done, the center area of the captured image is set as the in-focus area, so the camera optical axis of the imaging camera is controlled to be deflected up and down and left and right according to the running state of the host vehicle. The focal area can be changed. In addition, by controlling the aperture value of the imaging camera to create a so-called shallow depth-of-field state, and in that state, setting a focusing area focused on a predetermined area in front of the host vehicle, the focusing is performed. You may make it enlarge the amount of defocusing of area | regions other than.

  In the embodiment, the light distribution pattern of the headlamp HL is controlled based on the other vehicle detected by the vehicle detection device 1, but this is applied to a headlamp control device that controls the light distribution direction and the light intensity. You can also. Alternatively, the vehicle detection device 1 of the present invention can detect other vehicles existing not only in the front area of the own vehicle but also in the surrounding area, and can be applied to control of the vehicle speed and steering of the own vehicle based on this detection. is there.

  The present invention can be adopted as long as it is a vehicle detection device that images the front area of the host vehicle, detects a light spot from the captured image, and detects another vehicle based on the detected light spot.

DESCRIPTION OF SYMBOLS 1 Vehicle detection apparatus 2 Headlamp control apparatus (light distribution control part)
DESCRIPTION OF SYMBOLS 11 Focus area | region setting part 12 Light spot detection part 13 Other vehicle detection part 111 Focus pattern storage part 112 Focus processing part 113 Moving average calculating part 114 Selection part CAR Car (own vehicle)
HL Headlamp CAM Imaging camera P1 to P3 Focus pattern A1 to A3 Focus area LD (LDi, LDo) Light spot

Claims (4)

A vehicle detection device including other vehicle detection means for detecting a light spot in an image captured in the own vehicle and detecting another vehicle based on the light spot, wherein the predetermined area of the image is a focusing area, and the like Focusing area setting means for setting the area as the out-of-focus area, and the other vehicle detecting means detects a size of a clear light spot in the in-focus area and a blurred light spot in the non-focus area. When the detected size is smaller than the reference value, the light spot is detected as another vehicle. The vehicle detection apparatus according to claim 1, wherein the focus area setting unit changes the focus area according to a vehicle speed and a steering direction of the host vehicle. The focus area setting means sets, when the other vehicle detection means detects another vehicle, an area including the other vehicle closest to the host vehicle among the detected other vehicles as a focus area. The vehicle detection device according to claim 1. 2. The vehicle detection device according to claim 1, wherein the in-focus area setting unit sets the farthest forward position of the own vehicle in the in-focus area when the other vehicle detection unit does not detect the other vehicle .

Images