JP6899673B2 - Object distance detector - Google Patents

Description

The present invention relates to an object distance detecting device.

As a background technique in the present technical field, for example, Patent Document 1 proposes a method of suppressing an increase in the amount of calculation in distance detection by a stereo camera.

Specifically, Patent Document 1 describes that an image is reduced, stereo distance detection is performed on a reduced image with a reduced resolution, and stereo distance detection is performed by dividing an area with an image that does not reduce the resolution. ..

Japanese Unexamined Patent Publication No. 2015-230703

The technique described in Patent Document 1 can reduce the calculation load because the distance is detected on the image whose resolution is reduced by reducing the image, but the amount of information is reduced, so that the accuracy of the detected distance is high. It may decrease regardless of the distance of the subject. In addition, a processing circuit for reducing the image is required.

Therefore, an object of the present invention is to provide an object distance detection device capable of achieving both improvement in object detection accuracy and reduction in calculation load.

As an example, the object distance detection device of the present invention includes a plurality of imaging units and an image element corresponding to a specific image element in a reference image captured by one of the plurality of imaging units. The search is performed based on the search condition setting unit that sets the conditions for searching in the reference image captured by the other imaging unit and the conditions set by the search condition setting unit, and is obtained by the search. A stereo distance detection unit that detects the distance of an object based on a parallax is provided, and the stereo distance detection unit searches for a predetermined image area based on a first search condition set by the search condition setting unit. After that, the detected image region of the object is searched based on the second search condition set by changing the search density in the reference image with respect to the reference image.

The present invention can provide an object distance detection device that can achieve both improvement in object detection accuracy and reduction in calculation load.

It is a figure which shows the structure of the object distance detection apparatus in the Example of this invention. It is a figure which shows an example of the captured image imaged in the Example of this invention. It is a figure which shows an example of the captured image and area control imaged in the Example of this invention. It is a figure which shows an example of the method of specifying the search range specified in the Example of this invention. It is a figure which shows the processing timing of the object distance detection apparatus in the Example of this invention. It is a figure which shows the processing flow of the object distance detection apparatus in the Example of this invention. It is a figure which shows an example of the photographed image and the recognition result in the Example of this invention. It is a figure which shows an example of the current consumption of the stereo distance detection part in the Example of this invention. It is a figure which shows another example of area control in the Example of this invention. It is a figure which shows an example of the photographed image and the recognition result in the Example of this invention. It is a figure which shows the structure of the object distance detection apparatus in another Example of this invention.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a diagram showing a configuration of an embodiment of an object distance detecting device of the present invention. Reference numeral 1 denotes an object distance detecting device of the present embodiment, which is mounted in front of a vehicle, for example, and is configured as a part of a safety system that recognizes signals, signs, obstacles, and the like to assist the driver.

Reference numerals 101 and 102 are imaging units, and an optical lens is attached to the image sensor. These imaging units repeatedly capture one image at a predetermined timing and output the captured image. The imaging unit 101 and the imaging unit 102 are installed at a predetermined distance to the left and right, and the distance to the subject can be calculated from the deviation of the images captured by the imaging unit 101 and the imaging unit 102, that is, the so-called parallax.

Although FIG. 1 shows an example in which the components of the object distance detection device 1 are housed in the same housing, for example, the imaging units 101 and 102 are housed together in a housing different from the other components. Alternatively, the imaging units 101 and 102 may be housed in separate housings and attached to the vehicle. In that case, the image signal may be connected with a connection cable (not shown). As a method of transmitting an image using a connection cable, there is a transmission method using a differential transmission line of an LVDS (Low Voltage Differential Signaling) method.

Further, by using the image sensors of the imaging unit 101 and the imaging unit 102 as color image sensors, it is possible to acquire the color information of the captured image, and the brightness information of the traffic signal and the state of the taillight of the preceding vehicle can be obtained. In addition, it will be possible to judge from the color information.

Reference numeral 103 denotes an image correction unit, which captures images from the imaging units 101 and 102, corrects them with correction values measured in advance so as to match the brightness of each image, and further corrects image distortion by a lens and images. This is also corrected by a pre-measured correction value so as to align the horizontal positions of the images of units 101 and 102. The measurement of the correction value is performed in the manufacturing process of the object distance detection device. For each device before applying the correction value, the brightness correction value for each pixel that images a specific subject and makes the brightness of the acquired image uniform, and for each pixel that cancels the lens distortion and makes the image horizontal. The geometric correction value is obtained, and each device is stored as a correction table in, for example, a non-volatile memory (not shown). Further, in the captured image, the area to be corrected can be set by the coordinates according to the instruction from the processing area setting unit 106 described later.

Reference numeral 105 denotes a stereo distance detection unit, which detects the distance of the subject by inputting an image from the image correction unit 103. For example, there are the following methods for detecting the distance. The stereo distance detection unit 105 captures the image from the image correction unit 103 and calculates the parallax. As described above, since the imaging unit 101 and the imaging unit 102 are installed at a predetermined distance from each other to the left and right, the captured images have parallax. So-called stereo processing for calculating this parallax is performed.

As a parallax calculation method, for example, there is a block matching method. The stereo distance detection unit 105 calculates the parallax for, for example, the area of the image designated by the processing area setting unit 106, which will be described later, among the images from the image correction selection unit 103.

Specifically, first, on the image of the image pickup unit 102, which corresponds to a block area of a predetermined size, for example, 8 pixels in the vertical direction and 8 pixels in the horizontal direction, which is cut out from the designated image area of the image captured by the image pickup unit 101. Explore the area. On the image of the image pickup unit 102, the block area of the same size is shifted in the horizontal direction by the number of pixels designated as the search density by the search range setting unit 107 described later, and the correlation value at that time is evaluated. At that time, the search is performed for the search range corresponding to the number of pixels specified by the search range setting unit 107 as the search width in the horizontal direction. For example, by setting a combination in which the search range is 128 pixels and the search density is 2 pixels, or a combination in which the search range is 32 pixels and the search density is 1 pixel, the processing load of calculation including the designation of the processing area is set. And the accuracy can be controlled. The larger the search density, the coarser the accuracy of the distance, but the processing load in the search range is reduced, and the smaller the search density, the higher the accuracy of the detected distance.

Then, the difference in the positions of the matching block regions between the imaging unit 101 and the imaging unit 102 becomes the parallax represented by the number of pixels. Using this parallax, the distance of the object reflected in the block area in the real environment can be obtained. In this example, a block area is adopted as an image element for which the distance is to be obtained. As a method of matching comparison for evaluating the correlation value, for example, the position where the sum of the luminance differences of the pixels in the block region to be compared becomes small is taken as the parallax.

It is known that the detected distance can be obtained from the lens focal length of the imaging unit 101 and the imaging unit 102, the distance between the imaging unit 101 and the imaging unit 102, the parallax obtained above, and the pixel pitch of the imaging sensor. However, the method of calculating the distance is not limited to this. Further, the image element for which the distance is to be obtained is not limited to the above-mentioned block region, and individual pixels constituting the image sensor may be adopted. The method of detecting an object is, for example, when distance information indicating almost the same distance exists nearby, they are grouped as one set, and when the size of the group is a certain size or more, the group is regarded as an object. ..

Reference numeral 110 denotes a search condition setting unit, which is a specific image in a reference image captured by an image region corrected by the image correction unit 103 and a reference image captured by one of the plurality of imaging units described above by the stereo distance detection unit 105. Set the conditions for searching the image element corresponding to the element in the reference image captured by another imaging unit. Specifically, the search condition setting unit 110 includes a processing area setting unit 106 and a search range setting unit 107, which will be described later.

Reference numeral 106 denotes a processing area setting unit, which sets an image area for detecting a distance in the stereo distance detection unit 105, which is determined in advance or based on the result of the stereo distance detection unit 105 described later, and is also described later. An image area including an object is set in the recognition unit 108 for recognizing the object.

Reference numeral 107 denotes a search range setting unit, which is a search range for detecting a distance in the stereo distance detection unit 105 based on a predetermined or result of the stereo distance detection unit 105, and in a reference image with respect to a reference image. The search density, which is the amount of movement of the image element cut out in the above-mentioned block area in each pixel, is set.

Here, as the first search condition, for example, the image area is designated as almost the entire captured image, the search range is 128 pixels, and the search density is 2 pixels, which are predetermined values. As the second search condition, a limited area is set based on the result of the stereo distance detection unit 105, and the search range is set to 64 pixels and the search density is set to 1 pixel.

The stereo distance detection unit 105 performs the above-mentioned distance detection in the set area in the image from the correction unit 103 with the set search range and search density, and outputs the result to the recognition unit 108 described later. This makes it possible for the stereo distance detection unit 105 to determine the distance detection area, the search range, and the search density, respectively, and it is possible to avoid an increase in the processing load.

Reference numeral 108 denotes a recognition unit, which receives the detection result from the stereo distance detection unit 105 and the area designation from the processing area setting unit 106, recognizes the object on the image, and transmits the recognition result information to the outside of the object distance detection device 1. Output to. The object is recognized based on the distance information of the area specified by the processing area setting unit 106 obtained by the stereo distance detection unit 105. The method of recognizing an object is based on the size and shape of the detected group described above, and detects, for example, a vehicle or a pedestrian. There is a method of detecting the size and shape of an object by comparing it with pattern data held in advance as reference data. According to this processing method, since the distance of an object in front of the own vehicle, such as a pedestrian or a vehicle, can be obtained with high accuracy, it is used as information for avoiding a collision such as deceleration or stop of the own vehicle. In the above, as the detection result of the stereo distance detection unit 105, an example of detecting a distance group of a certain size as an object and outputting the result to the search condition setting unit 110 has been shown, but the recognition result of the recognition unit 108 is shown. It may be output to the search condition setting unit 110.

In the object distance detection device 1, for example, the image pickup units 101 and 102, the image correction unit 103, and the stereo distance detection unit 105 in the dotted line frame 12 are composed of electronic circuits, and other components are software by a microcomputer or the like (not shown). It is realized by processing. It is also possible to realize the stereo distance detection unit 105 by software processing.

FIG. 2 is a diagram showing an example of a captured image captured by an embodiment of the object distance detecting device of the present invention. In the figure, 1001 shows an image taken by the image pickup unit 101 and corrected by the correction unit 103, and 1002 shows an image taken by the image pickup unit 102 and corrected by the correction unit 103. 202, 203, and 204 are pedestrians, oncoming vehicles, and traffic signals that are the subjects.

Further, 201 and 209 indicate a common imaging region which is a region commonly captured in the captured image 1001 and the captured image 1002. As described above, there is a deviation in the region that is commonly captured between the captured image 1001 and the captured image 1002, and the distance of the subject is calculated from this deviation amount, that is, the parallax.

FIG. 3 is a diagram showing an example of a captured image captured by an embodiment of the object distance detecting device of the present invention and search condition setting. In the figure, 201 indicates, for example, a region of the captured image captured by the imaging unit 101 and corrected by the correction unit 103, which is captured in common with the image captured by the imaging unit 102 as described above. ..

205, 206, 207 and 208 are processing areas set by the processing area setting unit 106 in the captured image 201. The processing area 208 is a processing area set by the first search condition, and the processing areas 205, 206, and 207 are the results of the distance detection processing performed by the stereo distance detection unit 105 under the first search condition. It is a processing area set as a second search condition based on the above. That is, the stereo distance detection unit 105 detects the existence of an object and its distance in the processing areas 205, 206, and 207 in the processing areas 208 under the first search condition, and the processing areas 205 and 206 are used as the second search condition. , 207, the search range setting unit 107 finely sets the search density, so that the distance detection process of the object in the region including the pedestrian 202, the oncoming vehicle 203, and the traffic signal 204 is performed.

FIG. 4 is a diagram showing a relationship between a distance and a parallax according to an embodiment of the object distance detecting device of the present invention, and an example of a method of setting a search range by the search range setting unit 107.
The parallax is expressed in units of pixels, and indicates the amount of deviation of the image captured by the imaging unit 102 from the image captured by the imaging unit 101 described above. The distance and parallax are inversely proportional, and the parallax is large when the distance is short, and small when the distance is long. Therefore, the longer the distance, the larger the detection error and the lower the accuracy.

For example, when the area 205 included in the pedestrian 202 is detected by the stereo distance detection unit 105 to be about 5 m, the parallax at that time is 60 pixels from the figure. Therefore, as a second search condition, the search range setting unit 107 sets a range of 20 pixel widths of 51 pixels to 70 pixels with respect to the processing area 205 as a search range for the distance of the pedestrian 202. To do. At that time, the search density is set to 1 pixel. Similarly, when distances of about 20 m and about 50 m are detected with respect to the processing areas 207 and 206, which are the areas of the traffic signal 204 and the oncoming vehicle 203, the search range is 15 pixel width of 23 to 37 pixels. The 10 pixel widths of 1 pixel to 10 pixels are set respectively, and the search density is set to 1 pixel.

As a result, in each processing region, the parallax may be calculated in a limited search range centering on the distance detected by the stereo distance detection unit 105 under the first search condition, and the processing load is reduced. Further, the search densities may be different from each other, and by increasing the search density where the parallax value is large, the processing load can be further reduced without impairing the distance accuracy.

Further, when the distance cannot be detected in the set search range, the processing load can be minimized by expanding the set search range and performing the search process in the expanded range. Since the distance of the region 205 is short and the accuracy of detection can be ensured, the processing load is further reduced by removing it from the second search condition and limiting the region 205 to the region including a distant object as the second search condition. be able to.

FIG. 5 is a diagram showing processing timing of an embodiment of the object distance detecting device of the present invention. In the figure, (5-1) is the processing timing of the image correction unit 103 and the stereo processing detection unit 105 under the first search condition, and (5-2) is the image correction unit 103 under the second search condition. The processing timing of the stereo distance detection unit 105 is shown.

In (5-1), the above-mentioned distance detection process is performed by the image correction unit 103 and the stereo distance detection unit 105 under the first search condition. Further, in (5-2), regarding the processing areas 205, 206, and 207 specified by the processing area setting unit 106, which is the second search condition, the search range of each processing area specified by the search range setting unit 107 is obtained. Distance detection is performed by the stereo distance detection unit 105 at the specified search density. By processing in order from the area including the object closest to the own vehicle, the subsequent recognition unit 108 can recognize the object more quickly, and the safety can be ensured.

As described above, in the distance detection process in the stereo distance detection unit 105, the distance detection process is performed only for the designated required processing area with an appropriate search density and the minimum necessary search range in each processing area. It is not necessary to perform a detailed distance search for the entire range of the entire region of the captured image, and the processing load can be reduced.

FIG. 6 is a diagram showing a processing flow of an embodiment of the object distance detecting device of the present invention. First, images are taken by the imaging units 101 and 102 (S601: S represents a step). As described above, in each of the captured images, the processing area of the first search condition is set for the image correction unit 103 and the stereo distance detection unit 105 by the processing area setting unit 106 (S602), and the brightness is set by the image correction unit 103. Correction, lens distortion correction and horizontal alignment are performed (S603). Further, the search range setting unit 107 sets the search range and the search density of the first search condition for the stereo distance detection unit 105 (S604). Next, the stereo processing detection unit 105 detects an object in the set processing area and its distance (S605).

From the detection result, as a second search condition, the processing area setting unit 106 outputs the area information to be processed to the image correction unit 103 and the stereo distance detection unit 105 (S606), and the search range setting unit 107. From the distance information of the detected object, the search range and the search density to be searched in the area including the object are determined and output to the stereo distance detection unit 105 (S607). The image correction unit 103 corrects the image of the processing area according to the second search condition (S608).

Next, based on the obtained detection result, the stereo distance detection unit 105 detects a detailed distance in each search range in each designated area (S609).

Finally, the recognition unit 108 performs object recognition processing based on the distance detection result of the object in each processing area, and outputs the recognition result (S610). These processes are repeated, for example, every frame.

FIG. 7 is a diagram showing an example of a captured image and a recognition result in an embodiment of the object distance detecting device of the present invention.
Reference numeral 701 is an image captured by the imaging unit 101 at a certain point in time, and the imaging unit 102 also captures and acquires an substantially similar image. Further, 702, 703, and 704 are the recognition results of the object, and the frame and the distance display in the image are not shown in the captured image, but are explicitly shown on the image.

The pedestrian 202 detected from the processing area 205 is at a distance of 5.2 m, the oncoming vehicle 203 detected from the processing area 206 is at a distance of 53.1 m, and the traffic signal 204 at the processing area 207 is at a distance of 19.7 m. Is being detected. In this way, it is possible to realize highly accurate distance detection of an object over the entire captured image.

According to this embodiment, an object and its distance are roughly detected from the captured image, and based on the result, an accurate distance detection process is performed for a region including the object by limiting the search range of the distance. Object recognition of the whole becomes possible without increasing the processing load.

FIG. 8 is a schematic diagram of the current consumption waveform of the distance detection unit 105 during the distance detection operation in one embodiment of the object distance detection device of the present invention.
FIG. 8A is an operating current waveform when distance detection is performed by stereo processing over the entire region of the captured image, not based on the present invention, and power is consumed over the entire frame.
FIG. 8B is a current waveform during a distance detection operation in an embodiment of the object distance detection device of the present invention, and power is consumed at a limited timing within the frame with respect to FIG. 8A. Therefore, it is possible to reduce power consumption.

FIG. 9 is a diagram showing a configuration of another embodiment of the object distance detecting device of the present invention. The object distance detection device 1 is mounted on a vehicle such as an automobile, and in the figure, 801 is a vehicle control unit. The output of the recognition unit 108 is input to the vehicle control unit 801. Further, the recognition unit 108 includes a road surface detection unit 1081 that separates the road surface portion from other objects from the distance information in the captured image, detects it as a road surface region, and outputs the region to the search condition setting unit 110. Further, although not shown, the recognition unit 108 includes a travel route prediction unit 1082 that inputs vehicle information such as vehicle speed and steering angle, predicts the travel route of the own vehicle, and outputs the output to the search condition setting unit 110.

The search condition setting unit 110 inputs the road surface information detected by the road surface detection unit 1081 and sets the processing area in the image correction unit 103 and the stereo distance detection unit so as to detect the distance of the object on the road surface. As a result, the stereo distance detection unit 105 detects the exact distance of the object on the road surface and the recognition unit 108 recognizes it, so that the obstacle of the own vehicle can be recognized quickly and safe driving can be maintained. Further, the search condition setting unit 110 sets the processing area so that the travel route prediction unit 1082 detects the distance of the object on the route to be traveled, so that the search condition setting unit 110 sets the processing area on the travel route of the own vehicle as described above. It is possible to recognize obstacles and realize safe driving.

Further, although not shown, the search range setting unit 107 inputs the vehicle speed of the own vehicle, for example, when the speed is high, the search range is set wide, and when the speed is slow, the search range is set narrow, so that the processing amount is minimized. The distance detection process can be performed reliably.

Further, as shown in FIG. 10, for example, the processing area setting unit 106 sets the processing area 210, which is the entire image, as the processing area as the second search condition, and at that time, the search control unit uses 60 pixels as described above. By setting the range of 50 pixels to 70 pixels in the center as the search range and detecting the proximity distance of the own vehicle, for example, from 2 m to 10 m, the entire screen is the target of distance detection for the vicinity of the own vehicle. And can improve safety.

Further, the stereo distance detection unit 105 processes a plurality of adjacent objects together, for example, the processing area setting unit 106 treats them as one area, and the search range setting unit 107 detects and collects them by the stereo distance detection unit 105. By setting the search range including the distances of objects in one area, it is possible to deal with the case where a large number of objects exist.

The vehicle control unit 801 of FIG. 9 receives the recognition result by the recognition unit 108 and controls other devices (not shown) of the vehicle. Vehicle control includes lighting of warning lamps for drivers due to approaching pedestrians and detection of red lights and road signs, generation of warning sounds, deceleration by braking, stop control, and when following a preceding vehicle. Throttle, brake control, and other steering angle controls for collision avoidance and lane keeping. These vehicle control information is output from the object distance detection device 1 to other devices (not shown) via the in-vehicle network.

Note that FIG. 9 shows an example in which the vehicle control unit 801 is housed in the same housing as the object distance detection device 1, but the present invention is not limited to this, and as described above, the imaging units 101 and 102 are used as separate housings. May be good.

FIG. 11 is a diagram showing the configuration of still another embodiment of the object distance detecting device of the present invention. Reference numeral 901 is a network imaging unit, 903 is a LAN (Local Area Network), and 904 is a control unit. The network imaging unit 901 is connected to the control unit 904 via the LAN 903. Further, 902 is an image compression / interface unit, 905 is a network interface unit, and 906 is an image expansion unit.

The image captured by the imaging unit 101 and the imaging unit 102 is subjected to luminance correction, lens distortion correction, and horizontal alignment in the image correction unit 103. Next, the image compression / interface unit 902 compresses the image from the image correction unit 103 and transmits it to the LAN 903. As an image compression method, in order to reduce the processing time, there is a method using an in-screen compression method that compresses within one image without using the temporal correlation of a plurality of images. Further, the video compression coding method may be selected and switched.

The image compression / interface unit 902 generates compressed coded data and transmits the data according to a predetermined network protocol. The image correction unit 103 may be provided after the image expansion unit 906 of the control unit 904, but after the lens distortion and the like are corrected by processing in the front stage of the image compression / interface unit 902 of the network image pickup unit 901. Image compression is expected to improve the efficiency and image quality of image compression. In that case, the processing area setting by the processing area setting unit 106 is transmitted from the network interface unit 905 to the image compression / interface unit 902 and the image correction unit 103 via the LAN 903.

In the control unit 904, the network interface unit 905 receives the compressed image data via the LAN 903. The compressed image data received by the network interface unit 905 of the control unit 904 is expanded to the original image for the processing area set by the processing area setting unit 106 in the image expansion unit 906, and the distance is detected by the stereo distance detection unit 105. Will be done. The subsequent processing is as described above.

According to this embodiment, since the compressed image is transmitted via LAN906, the processing amount on the imaging unit side can be reduced, the weight on the imaging unit side is reduced, the power consumption is reduced, and the housing is reduced. It is possible to reduce the dimensional restrictions on vehicle installation.

The present invention is not limited to the above-described examples, and includes various modifications. For example, the above-described embodiment has been described in detail in order to explain the present invention in an easy-to-understand manner, and is not necessarily limited to the one including all the described configurations. Further, it is possible to replace a part of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add the configuration of another embodiment to the configuration of one embodiment. Further, it is possible to add / delete / replace a part of the configuration of each embodiment with another configuration.

In addition, each of the above configurations may be configured in part or in whole in hardware, or may be configured to be realized by executing a program in a processor. In addition, the control lines and information lines indicate those that are considered necessary for explanation, and do not necessarily indicate all the control lines and information lines in the product. In practice, it can be considered that almost all configurations are interconnected.

1 ... Object distance detection device, 101-102 ... Imaging unit, 103 ... Image correction unit, 105 ... Stereo distance detection unit, 106 ... Processing area setting unit, 107 ... Search range setting unit, 108 ... Recognition unit, 110 ... Search condition Setting unit, 201 ... Common imaging area, 202 to 204 ... Subject, 205 to 210 ... Processing area, 702 to 704 ... Recognition result, 801 ... Vehicle control unit, 901 ... Network imaging unit, 902 ... Image compression / interface unit, 903 ... LAN, 904 ... Control unit, 905 ... Network interface unit, 906 ... Image expansion unit.

Claims (11)

With multiple imaging units
Conditions for searching an image element corresponding to a specific image element in a reference image captured by one of the plurality of imaging units in a reference image captured by the other imaging unit. Search condition setting unit to set
A stereo distance detection unit that performs the search based on the conditions set by the search condition setting unit and detects the distance of the object based on the parallax obtained by the search is provided.
The stereo distance detection unit
After searching a predetermined image area based on the first search condition set by the search condition setting unit,
The detected image area of the object was, had row a search based on the second search condition set by changing the search density in said reference image with respect to the reference image,
The search condition setting unit is
A processing area setting unit that sets the search area in the reference image, and
A search range setting unit for setting a search range which is a search width for each pixel of the image element in the reference image with respect to the reference image and resetting the search density is provided.
The processing area setting unit sets a processing area including a plurality of objects obtained by the stereo distance detection unit.
The search range setting unit is an object distance detection device that sets the search range based on the distances of a plurality of the objects detected by the stereo distance detection unit. With multiple imaging units
Conditions for searching an image element corresponding to a specific image element in a reference image captured by one of the plurality of imaging units in a reference image captured by the other imaging unit. Search condition setting unit to set
A stereo distance detection unit that performs the search based on the conditions set by the search condition setting unit and detects the distance of the object based on the parallax obtained by the search is provided.
The stereo distance detection unit
After searching a predetermined image area based on the first search condition set by the search condition setting unit,
The detected image area of the object was, had row a search based on the second search condition set by changing the search density in said reference image with respect to the reference image,
The search condition setting unit is
A processing area setting unit that sets the search area in the reference image, and
A search range setting unit for setting a search range which is a search width for each pixel of the image element in the reference image with respect to the reference image and resetting the search density is provided.
When the stereo distance detection unit cannot detect the distance of the object, the search range setting unit expands the search range.
An object distance detection device that detects a distance with respect to the expanded search range in the stereo distance detection unit. In the object distance detecting device according to claim 1 or 2.
The object distance detection device, wherein the search density is a pixel-by-pixel movement step amount of the image element in the reference image with respect to the reference image. In the object distance detecting device according to claim 1 or 2.
The processing area setting unit sets one or a plurality of areas including the object acquired by the stereo distance detection unit based on the first search condition as a region for performing a search under the second search condition. , Object distance detector. In the object distance detecting device according to claim 1 or 2.
The search range setting unit sets the search density, which is finer than the first search condition, as the second search condition for the processing area set by the processing area setting unit. .. In the object distance detecting device according to claim 1 or 2.
The search range setting unit refers to the processing area set by the processing area setting unit according to the distance information of the object acquired by the stereo distance detection unit based on the first search condition. An object distance detection device that sets a search range smaller than the search condition of 1 as the second search condition. In the object distance detecting device according to claim 1 or 2.
An object distance detecting device including a geometric correction unit that geometrically corrects a predetermined area of an image captured by a plurality of the imaging units based on an area set by the processing area setting unit. In the object distance detecting device according to claim 1 or 2.
Equipped with a road surface detecting means for detecting the road surface
The processing area setting unit is an object distance detecting device that sets an area including a road surface detected by the road surface detecting means as a processing area. In the object distance detecting device according to claim 1 or 2.
Equipped with a travel route prediction means for predicting a travel route,
The processing area setting unit is an object distance detection device that sets an area including a route predicted by the traveling route prediction means as a processing area. In the object distance detecting device according to claim 1 or 2.
The search range setting unit is an object distance detection device that changes and sets the search range according to the vehicle speed. In the object distance detecting device according to claim 1 or 2.
The search range setting unit, a search range corresponding to the vicinity distance in the range of the search,
The processing area setting unit is an object distance detection device having the entire image area as a processing area.

Images