JP2016018371A - On-vehicle system, information processing apparatus, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To propose an on-vehicle system, an information processing apparatus, and a program capable of detecting a running available area with high accuracy.SOLUTION: A CPU 11 detects ranging points included in a predetermined height range on the basis of a detection signal of a laser radar 3, narrows down the ranging points by performing HOUGH transform, and performs projection transform on a position of each narrowed-down ranging point into coordinates in an image space of a picked-up image picked up by an on-vehicle camera 5. The CPU 11 rotates an edge filter. The CPU 11 calculates a gradient direction of the ranging point on the image from two-dimensional coordinates (x, y) of the ranging point obtained by the projection transform in the image space, rotates an edge detection filter along the calculated gradient direction, and sets a direction of the edge detection filter. The CPU 11 detects edges from the picked-up image using the set edge detection filter.SELECTED DRAWING: Figure 1

Description

  The present invention relates to an in-vehicle system used by being mounted on a vehicle traveling on a road.

  A technique for detecting a region in which a vehicle can travel from a captured image of a camera mounted on the vehicle has been proposed. For example, a predetermined cost is calculated at each point of the image from the luminance value in the local region of the image, the cost for each route is accumulated along each of the plurality of routes on the image, and the accumulated cost evaluation is based on the highest route. There is a device for detecting a travelable area (see Patent Document 1).

JP 2013-3800 JP

  In the apparatus of Patent Document 1, for example, in a situation where there are many buildings and installations around the road such as a city road, particularly when the measurement target is a long distance, a lot of noise around the road boundary is included, and the vehicle can travel. There is a problem that it is difficult to accurately detect the region.

  The objective of this invention is proposing the vehicle-mounted system, information processing apparatus, and program which can detect a driving | running | working area | region with high precision.

  The invention according to claim 1, which has been made to solve the above-described problem, includes an imaging unit (5), an edge detection unit (25), an object detection unit (3, 21), and a direction determination unit (24). And an in-vehicle system (1). The imaging means images the periphery of the vehicle. The edge detection unit detects an edge from the captured image captured by the imaging unit using an edge detection filter that detects an edge along a predetermined gradient direction.

  The object detection means detects the position of the object existing around the vehicle. The direction determining means determines the gradient direction described above based on the position of the object detected by the object detecting means. The edge detection unit detects an edge from the captured image using an edge detection filter that detects an edge along the gradient direction determined by the direction determination unit.

  In the vehicle-mounted system configured as described above, since the direction of the edge detection filter can be determined based on the position of the object detected by the object detection means, it is possible to perform edge detection in consideration of the presence of the object. . Therefore, it is possible to detect the edge by reducing the influence of noise, and it is possible to detect the travelable region with high accuracy by using the edge thus detected.

  Note that the above-described edge is a feature point on the captured image, for example, a location where the brightness of the image changes sharply. Moreover, although the target object mentioned above is solid objects, such as a curb stone, a guard rail, and a side groove, which are arranged along the direction of the lane, it is not limited to these.

  The invention described in claim 7 is an information processing apparatus (9) comprising an edge detection means (25), an object detection means (3, 21), and a direction determination means (24). The edge detection means detects edges using an edge detection filter that detects edges along a predetermined gradient direction from the captured image captured by the imaging means (5) that captures the periphery of the vehicle.

  The object detection means detects the position of the object existing around the vehicle. The direction determining means determines the gradient direction described above based on the position of the object detected by the object detecting means. The edge detection unit detects an edge from the captured image using an edge detection filter that detects an edge along the gradient direction determined by the direction determination unit.

The information processing apparatus configured as described above can constitute a part of the in-vehicle system according to claim 1.
By the way, although each means which comprises the information processing apparatus of the said Claim 7 may be implement | achieved by a hardware, in order to make a computer function as each means of an information processing apparatus as described in Claim 8. It may be realized by the program. With such a program, the computer can function as the information processing apparatus according to the seventh aspect.

  In addition, the code | symbol in the parenthesis described in this column and a claim shows the correspondence with the specific means as described in embodiment mentioned later as one aspect, Comprising: The technical scope of this invention is shown. It is not limited.

It is a block diagram which shows a vehicle-mounted system. It is a flowchart which shows the process sequence of a driving | running | working area | region determination process. It is a figure explaining the ranging point detected by a laser radar. It is an example of the captured image imaged with the vehicle-mounted camera. FIG. 6 is a diagram for explaining projection conversion of a ranging point, where (A) is a diagram showing a ranging point detected by a laser radar, (B) is a diagram showing a schematic transformation matrix, and (C ) Is a diagram in which ranging points are superimposed on the captured image. It is a figure explaining the setting method of ROI of a captured image. (A), (B) is a figure explaining the size setting method of an edge detection filter. It is an image figure explaining rotation of an edge detection filter. It is an example of an edge detection screen. It is a figure explaining the extraction method of a driving | running | working area | region, Comprising: (A) is a figure which superimposed the ranging point on the edge detection screen, (B) is the edge row | line | column and ranging point of the target object in (A). It is the figure which extracted and expanded. It is a figure which shows the captured image on which the image which shows a driving | running | working area | region displayed on a display was superimposed.

Embodiments of the present invention will be described below with reference to the drawings.
(1) Configuration of the in-vehicle system 1 The in-vehicle system 1 is a system used by being mounted on a vehicle such as an automobile. As shown in FIG. 1, a laser radar 3, an in-vehicle camera 5, a display 7, and information processing And a device 9.

  The in-vehicle system 1 is a system that detects a travelable area on a road. The travelable area here refers to a road where curbs, guardrails, etc. are arranged along the lane direction, where the curb or guardrail breaks (the part of the curbs, etc. that are continuously arranged are not partly located) ) Or curb cut-off, etc., which means the area where the vehicle can move sideways from the driving lane.

  The laser radar 3 is a device that irradiates laser light and acquires reflected light of the laser light. The reflected light detection signal is output to the information processing device 9. In the object detection unit 21 described later of the information processing device 9, the position of the object existing around the own vehicle around the front of the own vehicle, which is a vehicle on which the in-vehicle system 1 is mounted, is based on the detection signal. To calculate. That is, the position of the object around the vehicle is detected by the laser radar 3 and the object detection unit 21.

  The object here is a three-dimensional object located around the vehicle, and is arranged along the road. The laser radar 3 and the object detection unit 21 detect the position of an object having a height different from that of the traveling road surface of the vehicle, more specifically, a three-dimensional object having a height within a predetermined range.

  The in-vehicle camera 5 is an image pickup device that picks up an image of the periphery of the vehicle, in particular, the front side. For example, a known CCD image sensor or CMOS image sensor can be used. The in-vehicle camera 5 images the front of the vehicle at a predetermined time interval (1/15 s as an example) and outputs the captured image to the information processing device 9.

  The display 7 is a display unit such as a liquid crystal display that displays an image, and displays an image according to a signal input from the information processing apparatus 9. In the present embodiment, the display 7 superimposes and displays an image indicating the travelable area on the image captured by the in-vehicle camera 5.

  The information processing device 9 detects the travelable area based on the detection signal of the laser radar 3 and the captured image captured by the in-vehicle camera 5 and controls the display 7 to display the image.

  The information processing apparatus 9 includes a CPU 11, a ROM 12 that stores a program executed by the CPU 11, a RAM 13 that is used as a work area when the CPU 11 executes a program, a flash memory that can electrically rewrite data, an EEPROM, and the like. The computer system includes a nonvolatile memory 14 and the like, and executes a predetermined process by executing a program.

  The information processing device 9 also inputs travel information 15 from an ECU and various sensors mounted on the vehicle. The travel information 15 includes information such as a shift range, vehicle speed, steering angle, or yaw rate.

  The information processing device 9 functions as an object detection unit 21, a projective conversion unit 22, an ROI determination unit 23, a filter determination unit 24, an edge detection unit 25, a determination unit 26, and a result integration output unit 27.

(2) Processing by Information Processing Device 9 The travelable area determination processing executed by the information processing device 9 will be described based on the flowchart shown in FIG. For example, this process is started when the vehicle starts running (when the vehicle speed exceeds a predetermined threshold).

  The following S1 and S2 are processing of the object detection unit 21, S3 is processing of the projective transformation unit 22, S4 is processing of the ROI determination unit 23, and S5 and S6 are processing of the filter determination unit 24. S7 is processing of the edge detection unit 25, S8 is processing of the determination unit 26, and S9 is processing of the result integrated output unit 27.

  In S1, the CPU 11 detects a distance measuring point included in a predetermined height range. Here, first, based on the detection signal of the laser radar 3, the position where the three-dimensional object exists around the vehicle is detected. Hereinafter, this position is referred to as a distance measuring point.

  FIG. 3 is a diagram for explaining detected distance measuring points. The laser radar 3 is mounted on the vehicle 101 and detects a plurality of ranging points 103. Each of the detected distance measuring points 103 has information on distance and height with reference to the laser radar 3.

  The CPU 11 extracts a distance measuring point 103 of a three-dimensional object having a height in a predetermined height range from the plurality of detected distance measuring points 103. In FIG. 3, the ranging point 103c indicated by the broken line is excluded because it is outside the predetermined height range, and the ranging points 103a and 103b are extracted.

  In addition, the height range mentioned above in this embodiment is 15-25 cm, and this is equivalent to the height of the curb on the road. In the present embodiment, detection by the laser radar 3 is performed using a curb as an object, and the following processing is executed assuming that the distance measuring point is a curb.

  In addition, the height here is a height based on the traveling surface of the vehicle. For example, if the current road position where the vehicle is traveling is horizontal but the road ahead is different, such as an uphill road, the current road position is the reference. If the height is measured as such, the height of the distance measuring point may not be an accurate value.

  Therefore, the numerical value may be corrected so that the height becomes the height from the road surface. For example, the height of the road in the vicinity of the distance measuring point (height based on the current vehicle position) is also measured, and the height of the distance measuring point is corrected so that the height of the road becomes the reference. It is possible.

  In S2, the CPU 11 narrows down the distance measuring points. Here, ranging points other than the curb are deleted from the candidate points (ranging points 103a and 103b) detected in S1. In many cases, the curbs arranged at the road boundary are linear or gently curved. Using this, HOUGH conversion, which is a well-known technique, is performed on the candidate points to calculate a straight line or a curve, and candidate points existing within a predetermined distance from the line are extracted as ranging points indicating curbstones. To do.

In FIG. 3, a distance measuring point 103 along the straight line 105 calculated by the HOUGH conversion is extracted as a distance measuring point indicating a curb.
In S2, the distance measuring points 107 detected at the previous time (those narrowed down in S2 executed in the past) are used for narrowing down by HOUGH conversion. The past distance measuring point 107 can correct the error associated with the traveling of the vehicle using the traveling information 15 such as the vehicle speed and the steering angle, but if the time difference between the current time and the previous time is sufficiently small, the correction is made. It does not have to be done. The distance measuring points narrowed down in S2 are stored in the RAM 13 for use in the subsequent narrowing down in S2.

  In S <b> 3, the CPU 11 performs projective conversion of the position of the distance measuring point narrowed down in S <b> 2 to the coordinates of the image space of the captured image captured by the in-vehicle camera 5. An example of the captured image is shown in FIG. In the captured image 111, the curb 113 is arranged along the lane, but a cut-down region 115 of the curb 113 is formed in part. This process executes a process for recognizing the cut-down area 115 as a travelable area.

  In S3, first, a captured image newly captured by the in-vehicle camera 5 is acquired. Next, the coordinates (X, Y, Z) of each ranging point 103a shown in FIG. 5A are imaged as shown in FIG. 5C using a transformation matrix as shown in FIG. 5B. A distance measuring point 117 indicated by coordinates (x, y, 1) on the image 111 is converted. Hereinafter, in the description of the captured image, the converted position of the distance measuring point 117 is simply referred to as the position of the distance measuring point.

  The captured image 111 can determine the range to be captured based on the installation position of the in-vehicle camera 5. Therefore, it is possible to superimpose a distance measuring point by the laser radar 3 on the image using the above-described conversion matrix.

  In S4, the CPU 11 sets an ROI (Region of Interest) of the captured image. The ROI is an area for edge detection. Here, as shown in FIG. 6, the ROI 121 is set so as to include the distance measuring point 117 in consideration of the position of the distance measuring point 117 in the captured image 111. The ROI 121 is set with a width around the position of the distance measuring point, but the width is set smaller as the distance from the vehicle to the distance measuring point 117 is longer.

  In S5, the CPU 11 sets the size of the edge detection filter for detecting the edge from the captured image. Here, first, the distance from the position of the laser radar 3 to the distance measuring point, in other words, the distance from the position of the own vehicle, from the information (X, Y, Z) of the distance measuring point narrowed down in S2. The real space distance to the point is calculated.

  Then, as shown in FIG. 7A, the CPU 11 sets the filter size so that the size of the edge detection filter 123 increases as the real space distance to the distance measuring point 117 is shorter. Since the real space distance is different for each ranging point 117, the edge detection filter 123 having a different size depending on the position in the ROI 121 is set.

  Note that when there is a distance measuring point in the vicinity of the vehicle as shown in FIG. 7A, an edge detection filter 123 having a small size to a large size is used, but as shown in a captured image 111a shown in FIG. 7B. When the distance between the vehicle and the curb stone 113 is large, a large edge detection filter is not used. In this way, a filter of an appropriate size is set according to the distance from the host vehicle.

  In S6, the CPU 11 rotates the edge filter. As described above, it is assumed that the curb lines installed on the road boundary are linear, and as shown in FIG. 6, the edge gradient direction is also limited to a limited direction.

In S6, the gradient direction on the image of the distance measurement point is calculated from the two-dimensional coordinates (x _i , y _i ) of the distance measurement point converted into the image space by the following formula 1, and along the calculated gradient direction. In other words, the edge detection filter 123 is rotated to set its direction so as to detect an edge along the gradient direction. By using the edge detection filter 123 rotated in this way, the influence of noise can be reduced and the edge of the curb can be detected with high accuracy.

  In the above equation 1, x is the x coordinate of the distance measuring point in the image space, and y is the y coordinate of the distance measuring point in the image space. N is the number of distance measuring points. Then, as shown in FIG. 8, the edge detection filter is rotated according to the calculated tan (θ). FIG. 8 shows an image of filter rotation.

  Note that the rotation and setting of the filter includes selecting and using a filter to be used from among edge detection filters that detect edges along various gradients in advance.

  In S7, the CPU 11 detects an edge from the captured image. Here, in the ROI set in S4 in the captured image, the edge is detected using the edge detection filter set in S5 and S6. FIG. 9 shows an example of the edge detection screen 125 as a result of edge detection. As in the edge detection screen 125, only edges along the direction of the edge detection filter in the ROI of the captured image are extracted.

In S8, the CPU 11 adds the curb distance measuring point by the laser radar 3 and the edge extracted from the captured image, and extracts the travelable area from the result.
In the present embodiment, first, the CPU 11 specifies an edge corresponding to a distance measuring point of the curb by the laser radar 3. As shown in FIG. 10A, when the distance measuring point 117 is superimposed on the edge detection screen 125 and the edge and the distance measuring point are overlapped, the edge row 131 corresponding to the distance measuring point 117, that is, the edge of the object. Can be identified. The CPU 11 uses this edge row 131 to complement the distance measuring points 117 for extracting the travelable area.

  As shown in FIG. 10B in which the edge row 131 and the distance measuring points 117 of the object in FIG. 10A are extracted and enlarged, the distance measuring points 117 indicating the curb do not exist, and the edge rows 131 further exist. The area that does not exist is determined to be the travelable area 133. On the other hand, if the edge row 131 extends so as to connect the distance measuring points 117, even if the distance between the distance measuring points 117 is large, it is determined that the travel impossible area 135 is in between.

  It should be noted that the area where only the distance measuring point 117 exists is determined to be a non-travelable area because some solid object exists even if no edge is detected. Further, an area where there is no distance measuring point 117 and only an edge can be assumed to be an area where the color on the road surface such as a small step or a white line is different, so that it is determined as a travelable area.

  In S9, the CPU 11 causes the display 7 to display a captured image 143 in which an image 141 indicating the travelable area is superimposed and displayed as shown in FIG. After S9, the process returns to S1. Therefore, the captured image 143 is continuously displayed while the vehicle is traveling.

(3) Effect In the in-vehicle system 1 of the present embodiment, the laser radar 3 and the object detection unit 21 detect the position of the object around the vehicle as a distance measuring point. The in-vehicle camera 5 captures the periphery of the vehicle, and the edge detection unit 25 detects the edge from the captured image. The edge detection unit 25 detects an edge using an edge detection filter. The filter determination unit 24 determines the gradient direction of the edge detection filter based on the distance measuring point of the object.

  That is, the in-vehicle system 1 of the present embodiment includes an imaging unit (5), an edge detection unit (25), an object detection unit (3, 21), and a direction determination unit (11, 24). The imaging means images the periphery of the vehicle. The edge detection unit detects an edge from the captured image captured by the imaging unit using an edge detection filter that detects an edge along a predetermined gradient direction.

  The object detection means detects the position of the object existing around the vehicle. The direction determination unit determines the above-described gradient direction of the edge detection filter used by the edge detection unit based on the position of the target detected by the target detection unit. The edge detection unit detects an edge from the captured image using an edge detection filter that detects an edge along the direction determined by the direction determination unit.

  In the vehicle-mounted system 1 according to the present embodiment configured as described above, the direction of the edge detection filter is determined along the direction in which the distance measurement points are arranged in the captured image based on the distance measurement points detected by using the curb as an object. Therefore, when detecting an edge from a captured image, the edge of the curb can be detected with high accuracy. Therefore, it is possible to detect a useful edge that is not easily affected by noise, and it is possible to detect a travelable region with high accuracy.

  Note that the position of the object detected by the object detection means is information capable of grasping at least the positional relationship with the vehicle. By grasping the positional relationship between the target object and the vehicle, the position where the target object exists in the captured image can be obtained, and thereby an appropriate direction of the edge detection filter can be determined.

An apparatus other than the laser radar may be used as an apparatus for detecting the presence of an object around the vehicle. For example, a millimeter wave radar can be used.
Further, in the in-vehicle system 1 of the present embodiment, the projective conversion unit 22 performs projective conversion of the distance measuring point of the object into coordinates on the captured image, and determines the above-described direction of the edge detection filter based on the converted position.

  That is, the in-vehicle system 1 of this embodiment includes a conversion unit (22) that converts the position of the object detected by the object detection unit into a position on the captured image, and the direction determination unit is converted by the conversion unit. The direction of the edge detection filter is determined based on the position of the object on the captured image.

  Since the position of the ranging point where the object is detected on the captured image is a position where the edge of the object is highly likely to be detected, it is based on the ranging point projected onto the captured image as described above. By determining the direction of the edge detection filter, the edge can be detected with high accuracy.

  Moreover, the target object detection part 21 of the vehicle-mounted system 1 of this embodiment detects the position of the target object whose height is the range of 15-25 cm in S1 of FIG. Therefore, the object detected by the laser radar 3 can be narrowed down to the curbstone, and the curbstone can be detected with high accuracy.

  In the present embodiment, the curb is set as an object and the travelable area is detected as an example. However, a structure other than the curb may be detected as the object. For example, by setting a range of height to be detected, a side groove, a guard rail, or the like can be set as an object.

  The processing of the object detection unit 21 and the edge detection unit 25 may be changed according to the object to be detected. For example, when a guardrail is used as an object, when detecting a guardrail from a captured image, the guardrail can be detected using parameters specific to the guardrail such as light reflection intensity, color, and brightness. The accuracy of detecting the edge of the guardrail can be increased.

  Further, if the detection accuracy from the captured image can be improved in this way, the edge and the distance measuring point can be detected even if the threshold for narrowing down the object by the laser radar 3 is lowered and the height range to be detected is set wider. By combining them, highly accurate object detection can be realized.

  Further, the determination unit 26 of the in-vehicle system 1 of the present embodiment uses the distance measurement points detected by the laser radar 3 and the object detection unit 21 in S8 of FIG. 2 and the captured image of the in-vehicle camera 5 by the edge detection unit 25. The extracted edges are added together, and it is determined from the result whether or not it is a travelable region.

  That is, the in-vehicle system 1 according to the present embodiment determines the region in which the vehicle can travel based on the edge detected by the edge detection unit and the position of the target detected by the target detection unit. Means (11, 26) are provided.

The in-vehicle system 1 configured as described above can realize detection of a travelable region with high accuracy by combining detection results of both the laser radar 3 and the in-vehicle camera 5.
In the present embodiment, the determination unit 26 determines that a region where no edge corresponding to the curb distance measurement point exists and no curb distance measurement point exists is the travelable area.

  That is, in the in-vehicle system 1 of the present embodiment, the determination unit corresponds to an area on the road where the object is not detected by the object detection unit, and the edge detected by the edge detection unit corresponds to the area on the captured image. In the case where the vehicle does not exist at the position where the vehicle travels, the region is determined as a region where the vehicle can travel.

  Since the in-vehicle system 1 configured as described above determines that no curb exists from the detection signals of both the laser radar 3 and the in-vehicle camera 5, it is possible to improve the accuracy of determining the presence of the curb. Therefore, it is possible to detect not only curb stones but also curb cut-downs.

  Note that if the edge described above extends so as to connect a plurality of distance measuring points, the determining unit 26 determines that the distance between the distance measuring points is an untravelable region during that time. Therefore, the in-vehicle system 1 can reduce the risk of being erroneously determined as the travelable area when the distance measuring point is spaced for some reason although there is a curb.

  In addition, the edge detection unit 25 in the in-vehicle system 1 of the present embodiment is a range in which an edge is detected from a captured image based on the position where the object is detected by the laser radar 3 and the object detection unit 21 in S4 of FIG. (ROI) is set.

  Since the in-vehicle system 1 configured in this way can narrow down the calculation range for performing image processing for detecting edges, the calculation cost can be reduced. Furthermore, since the in-vehicle system 1 sets the ROI to be smaller as the point on the captured image corresponding to the point far from the vehicle is, the noise component included in the ROI can be reduced.

  Further, the filter determination unit 24 in the in-vehicle system 1 of the present embodiment sets the filter size so that the edge detection filter having a larger size is used as the position where the object is detected by the laser radar 3 and the object detection unit 21 is closer. To do. In a region far from the vehicle, the noise included can be reduced by reducing the filter. On the other hand, in an area where the distance is short, the accuracy of detecting a necessary edge can be increased by increasing the filter.

[Modification]
Although the embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments, and it goes without saying that various forms can be taken as long as they belong to the technical scope of the present invention.

  For example, the specific processing content in the travelable area determination processing described in the above embodiment is merely an example, and is not limited to the above-described content. For example, in S2, the curb distance measurement points are extracted using HOUGH conversion, but the distance measurement points may be extracted by other methods. For example, it is conceivable to predict the position of the curb from past distance measuring points and extract only distance measuring points close to the predicted position.

The ROI setting and the filter size setting executed in S4 and S5 may be configured such that either one or both are not executed.
Further, in the above embodiment, as shown in FIGS. 10A and 10B, the edge row 131 along the distance measuring point 117 is extracted, and the edge row 131 is used complementarily based on the distance measuring point 117. Although the configuration for detecting the travelable area 133 has been illustrated, the present invention is not limited to this method, and the travelable area can be detected by various methods combining a distance measuring point and an edge.

Moreover, in the said embodiment, although the structure which displays the detected driveable area | region on the display 7 was illustrated, the usage method of the detected driveable area | region is not limited to it.
For example, when the vehicle has a function of performing automatic driving, the vehicle may be configured to detect a travelable region and travel in that region. In addition, if the vehicle has a function of issuing an alarm when it is predicted that the vehicle will deviate from the traveling lane, the vehicle may be configured not to determine that the vehicle has deviated from the lane when traveling to the travelable region. Good.

In addition, the function as each means with which the information processing apparatus 9 mentioned above is provided can be realized in a computer by a program.
The program consists of an ordered sequence of instructions suitable for processing by a computer, stored in a ROM or RAM incorporated in the computer, and may be loaded and used from there. It may be loaded into a computer and used via various recording media and communication lines.

Examples of the recording medium include optical disks such as CD-ROM and DVD-ROM, magnetic disks, and semiconductor memories.
The information processing apparatus 9 may be a programmable logic device such as ASIC (Application Specific Integrated Circuits) or FPGA (Field Programmable Gate Array), or a discrete circuit.

DESCRIPTION OF SYMBOLS 1 ... In-vehicle system, 3 ... Laser radar, 5 ... In-vehicle camera, 9 ... Information processing apparatus, 11 ... CPU, 21 ... Object detection part, 22 ... Projection conversion part, 23 ... ROI determination part, 24 ... Filter determination part, 25: Edge detection unit, 26: Determination unit

Claims (8)

Imaging means (5) for imaging the periphery of the vehicle;
Edge detection means (25) for detecting an edge from an image captured by the imaging means using an edge detection filter for detecting an edge along a predetermined gradient direction;
Object detection means (3, 21) for detecting the position of an object existing around the vehicle;
Direction determining means (24) for determining the gradient direction based on the position of the object detected by the object detecting means,
The in-vehicle system, wherein the edge detection unit detects an edge from the captured image using an edge detection filter that detects an edge along the gradient direction determined by the direction determination unit. Conversion means (22) for converting the position of the object detected by the object detection means into a position on the captured image;
2. The vehicle-mounted device according to claim 1, wherein the direction determination unit determines the gradient direction of the edge detection filter based on a position on the captured image of the object converted by the conversion unit. system. The in-vehicle system according to claim 1, wherein the object detection unit detects a position of the object having a height within a predetermined range. And a determination unit that determines a region in which the vehicle can travel based on the edge detected by the edge detection unit and the position of the target detected by the target detection unit. The in-vehicle system according to any one of claims 1 to 3. In the case where the determination unit does not include an edge detected by the edge detection unit in a position corresponding to the region on the captured image in an area on the road where the object is not detected by the object detection unit. The vehicle-mounted system according to claim 4, wherein the region is determined as a region where the vehicle can travel. The said edge detection means sets the range which detects an edge from the said captured image based on the position where the said target object was detected by the said target object detection means. Any one of Claims 1-5 characterized by the above-mentioned. The in-vehicle system according to claim 1. Edge detection means (25) for detecting an edge using an edge detection filter for detecting an edge along a predetermined gradient direction from a captured image captured by an imaging means (5) for capturing the periphery of the vehicle;
Object detection means (3, 21) for detecting the position of an object existing around the vehicle;
Direction determining means (24) for determining the gradient direction based on the position of the object detected by the object detecting means,
The information processing apparatus according to claim 1, wherein the edge detection unit detects an edge from the captured image using an edge detection filter that detects an edge along the gradient direction determined by the direction determination unit.   A program for causing a computer to function as each unit of the information processing apparatus according to claim 7.

Images