Vehicle control system

Abstract

A vehicle control system is provided with: a projection device that projects a predetermined pattern on a road surface in the vicinity of a host vehicle; a target recognition device that recognizes a specific target located in the periphery of the own vehicle on the basis of a peripheral image obtained by capturing an image of a peripheral region of the own vehicle, and outputs position information, which is information regarding the relative position of the own vehicle and the specific target; and a processor that controls the projection device and the object recognition device. The processor superimposes a graphic image projected on the road surface by the projection device on a target recognition image, which is an image of a region recognized as the specific target in the peripheral image. A predetermined correction process including correcting the relative position acquired on the basis of the position information is executed.

Description

Technical Field

[0001] The present invention relates to a vehicle control system including a projection device and an object recognition device. The projection device projects (casts) a predetermined pattern on the road surface around the host vehicle, and the object recognition device detects the position of a specific object (the distance between the host vehicle and the specific object) located around the host vehicle based on an image obtained by photographing the surrounding area of the host vehicle. Background Art

[0002] A projection device that projects a predetermined pattern on the road surface around the host vehicle has been proposed (for example, refer to Japanese Unexamined Patent Application Publication No. 2021-127071). Summary of the Invention

[0003] The following object recognition device is known: analyzing a surrounding image obtained by photographing the surrounding area of the host vehicle (based on learned data), recognizing a specific object (such as a pedestrian), and obtaining the position of the object (the distance between the host vehicle and the specific object). When a specific object enters the pattern projected on the road surface or is located near the pattern, the light beam (direct light or reflected light) of the projection device irradiates a part of the specific object. As a result, in the surrounding image, a part of the image of the specific object sometimes becomes blurred. In this case, in the surrounding image, the accuracy of the area recognized as the specific object by the object recognition device decreases, and the accuracy (accuracy) of the position information obtained based on the coordinates of the area decreases.

[0004] One object of the present invention is to provide a vehicle control system capable of suppressing a decrease in the detection accuracy of the position of a specific object. The vehicle control system includes a projection device and an object recognition device. The projection device projects a predetermined pattern on the road surface around the host vehicle, and the object recognition device detects the position of a specific object (the distance between the host vehicle and the specific object) located in the surrounding area of the host vehicle based on an image of the surrounding area of the host vehicle.

[0005] To solve the above problems, the vehicle control system of the present invention includes: a projection device that projects a predetermined pattern on the road surface around the host vehicle; an object recognition device that recognizes a specific object located around the host vehicle based on a surrounding image obtained by photographing the surrounding area of the host vehicle and outputs position information indicating the relative position between the host vehicle and the specific object; and a processor that controls the projection device and the object recognition device. The processor is configured to perform a predetermined correction process including correcting the relative position obtained based on the position information when an object recognition image, which is an image of an area recognized as the specific object in the surrounding image, overlaps with a pattern image projected on the road surface by the projection device.

[0006] The vehicle control system according to the present invention includes a projection device that projects a predetermined pattern on the road surface around the host vehicle, and an object recognition device that obtains position information indicating the relative position between the host vehicle and a specific object based on a surrounding image. Here, when the object recognition image (the area recognized as a specific object in the surrounding image) overlaps with the pattern image, the accuracy of the position information may sometimes be low. According to the vehicle control system of the present invention, when the two images overlap, the relative position of the specific object with respect to the host vehicle obtained based on the position information is corrected by a predetermined correction process. Thereby, a decrease in the detection accuracy of the position of the specific object with respect to the host vehicle is suppressed.

[0007] In the vehicle control system according to one aspect of the present invention, there is provided a position deviation map that defines a deviation amount between a first position and a second position, the first position and the second position being obtained respectively based on the position information output from the object recognition device in the same position relationship between the host vehicle and the specific object, the first position being the position of the specific object with respect to the host vehicle obtained based on the position information output when the pattern image overlaps with the object recognition image in a first state, the second position being the position of the specific object with respect to the host vehicle obtained based on the position information output when the pattern image does not overlap with the object recognition image in a second state, and the correction process includes: correcting the first position obtained in the first state based on the position deviation map.

[0008] Accordingly, in the first state, the processor can relatively simply correct the position of the specific object with respect to the host vehicle.

[0009] In the vehicle control system according to another aspect of the present invention, the position deviation map includes a plurality of position deviation tables selected based on at least one of the image size of the object recognition image, the brightness of the surrounding image, and the first position.

[0010] The accuracy of the position information is affected by the brightness of the surrounding image, the image size of the object recognition image, and the first position. In particular, in the first state, the influence of these conditions on the accuracy of the position information is large. According to the vehicle control system of the present aspect, the position of the specific object with respect to the host vehicle can be corrected based on at least one of these conditions.

[0011] In the vehicle control system according to another aspect of the present invention, the processor obtains the first position, the brightness of the surrounding image, and the image size of the target recognition image. The position deviation map includes a plurality of position-divided tables selected according to the first position. Each position-divided table includes a plurality of brightness-divided tables selected according to the brightness of the surrounding image. Each brightness-divided table includes a plurality of size-divided tables that define the deviation amount selected according to the image size.

[0012] According to the vehicle control system according to this aspect of the technology, it is possible to correct the position of a specific target relative to its own vehicle based on the first position, the brightness of the surrounding image, and the image size of the specific target.

[0013] In the vehicle control system according to another aspect of the present invention, the size-divided table defines the relationship between the image size and the deviation amount such that the larger the image size of the target recognition image, the smaller the deviation amount.

[0014] In the vehicle control system according to another aspect of the present invention, the surrounding image is an image obtained by photographing the foreground of its own vehicle. The target recognition device applies the surrounding image to a pre-learned deep neural network to recognize the specific target, and based on the coordinates of the lower end of the region recognized as the specific target in the surrounding image, obtains the distance in the vehicle longitudinal direction between its own vehicle and the specific target as the position information.

[0015] In the vehicle control system according to another aspect of the present invention, the target recognition image is a rectangular image, and the processor performs the correction process when the graphic image overlaps the lower end of the target recognition image.

[0016] When the lower end of the target recognition image overlaps the graphic image, the lower end of the target recognition image becomes blurred. Therefore, the accuracy of the position information output from the target recognition device is likely to be low. For example, the distance between the own vehicle and the specific target may be larger than the actual distance. According to the vehicle control system according to this aspect of the technology, when the lower end of the target recognition image overlaps the graphic image, the distance obtained based on the image (the distance between the own vehicle and the specific target) is corrected. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Hereinafter, the features, advantages, and technical and industrial significance of the exemplary embodiments of the present invention will be described with reference to the drawings. In the drawings, the same reference numerals denote the same elements, and:

[0018] Figure 1It is a block diagram of a vehicle control system related to an embodiment of the present invention;

[0019] Figure 2 It is an example of a foreground image;

[0020] Figure 3 It is a mapping that defines the relationship between a specified distance level, a brightness level, a longitudinal dimension level, and an offset value;

[0021] Figure 4 It is a graph showing the distance ΔD1 obtained when a pattern graphic is projected onto the road surface and the distance ΔD2 obtained when the pattern graphic is not projected;

[0022] Figure 5 It is a flowchart of a program executed by the CPU to achieve the function of correcting the distance ΔD obtained by the target recognition device. Detailed Embodiment

[0023] Overview

[0024] As Figure 1 shown, a vehicle control system 1 related to an embodiment of the present invention is applied to a vehicle V having an autonomous driving function (hereinafter referred to as "own vehicle"). In addition, the vehicle control system 1 has the following first notification function: project a predetermined pattern graphic on the road surface in front of the own vehicle and provide predetermined information to others located around the own vehicle. The vehicle control system 1 has the following second notification function: in a state where the autonomous driving function is invalidated, based on an image obtained by photographing the front area of the own vehicle, when the risk of contact between the own vehicle and a specific target object OB is high, provide predetermined information to the driver of the own vehicle.

[0025] Specific Configuration

[0026] As Figure 1 shown, the vehicle control system 1 includes an ECU 10, a projection device 20, a target recognition device 30, and a notification device 40.

[0027] The ECU 10 includes a microcomputer including a CPU 10a, a ROM 10b (rewritable non-volatile memory), a RAM 10c, a timer 10d, etc. The CPU realizes various functions by executing programs (instructions) stored in the ROM. The ECU 10 is connected to the ECUs of other devices via a CAN (Controller Area Network).

[0028] The projection device 20 irradiates the road surface in front of the host vehicle (right front and / or left front) with a light beam representing a pattern graphic corresponding to an instruction obtained from the ECU 10, and projects the pattern graphic onto the road surface.

[0029] The target object recognition device 30 includes an imaging device. The imaging device has, for example, a CCD built therein. The imaging device is provided at the front part of the host vehicle. The imaging device faces the front of the host vehicle. The imaging device captures the foreground of the host vehicle at a predetermined frame rate and obtains image data representing the foreground image PIC. The target object recognition device 30 further includes an image analysis device. The image analysis device obtains the image data from the imaging device and obtains the brightness BR (average value of the brightness of all pixels) of the foreground image PIC. In addition, the image analysis device analyzes the image data (applies it to a pre-learned deep neural network DNN (= Deep Neural Network)) to recognize the image of a specific target object OB (for example, a pedestrian) in the foreground image PIC and the image of the pattern graphic (pattern graphic image PTN) projected onto the road surface by the projection device 20 (see Figure 2 ). The image analysis device calculates the distance ΔD in the front-rear direction between the host vehicle and the specific target object OB based on the ordinate Y of the lower end line L1 of the image (target object recognition image) of the area (rectangular area R) recognized as the specific target object OB in the foreground image PIC. Specifically, the image analysis device obtains the distance ΔD with reference to a map (not shown) that defines the relationship between the ordinate Y and the distance ΔD. This map is designed such that the smaller the ordinate Y (the closer the rectangular area R is to the lower end of the foreground image PIC), the smaller the distance ΔD. The distance ΔD corresponds to information (position information) regarding the relative position between the host vehicle and the specific target object OB. In addition, the image analysis device obtains the size (longitudinal dimension H) of the rectangular area R. The longitudinal dimension H corresponds to the image size of the target object recognition image. The image analysis device provides these calculation results (brightness BR, longitudinal dimension H, and distance ΔD) to the ECU 10. In addition, the image analysis device determines whether the lower end line L1 of the rectangular area R overlaps with the pattern graphic image PTN in the foreground image PIC. The image analysis device provides this determination result to the ECU 10.

[0030] The notification device 40 includes an image display device and an audio device. The image display device is arranged, for example, on the instrument panel (near the speed display device). The image display device displays an image according to an instruction obtained from the ECU 10. The audio device reproduces a sound according to an instruction obtained from the ECU 10.

[0031] First notification function

[0032] When a predetermined condition is satisfied, the ECU 10 causes the projection device 20 to project a predetermined pattern graphic onto the road surface. For example, when the host vehicle turns left (or right) (when the direction indicator on the left (right) side of the host vehicle is operating (flashing)), the ECU 10 causes the projection device 20 to project an arrow pointing left (right) onto the road surface in front of the host vehicle at an oblique left (right) angle. Thereby, the traveling direction of the host vehicle is notified to pedestrians and drivers of other vehicles in the vicinity of the host vehicle.

[0033] Second notification function

[0034] The ECU 10 sequentially obtains the distance ΔD from the target recognition device 30. The ECU 10 calculates the average distance ΔDave, which is the average of the distances ΔD corresponding to a plurality of consecutive frames. When the average distance ΔDave is equal to or less than the threshold value ΔDth, the ECU 10 sends an instruction to the image display device of the notification device 40 to display a predetermined image (icon), and sends an instruction to the audio device of the notification device 40 to reproduce a predetermined sound (beep) to provide the driver of the host vehicle with information indicating a high risk of contact between the host vehicle and a specific target OB.

[0035] Here, when a specific target OB enters the pattern graphic projected onto the road surface, or when the specific target OB is located near the pattern graphic, the light beam (direct light or reflected light) of the projection device 20 irradiates a part of the specific target OB. As a result, in the foreground image PIC, sometimes a part of the image of the specific target OB becomes blurred. In this case, in the foreground image PIC, the accuracy of the region (rectangular region R) recognized as the specific target OB by the image analysis device decreases, and the accuracy (accuracy) of the distance ΔD obtained based on the coordinates of the lower end line L1 of the rectangular region decreases. Specifically, the distance ΔD obtained in a state where the lower end line L1 of the rectangular region R overlaps the pattern graphic image PTN (first state) is larger than the distance ΔD obtained in a state where the pattern graphic is not projected onto the road surface or the two do not overlap (second state) (it is necessary to confirm whether it is correct).

[0036] In addition, when the specific target object OB is far from the host vehicle, the image of the specific target object OB in the foreground image PIC becomes blurred, so the accuracy of the distance ΔD decreases. In addition, the smaller the brightness BR (surrounding brightness) of the foreground image PIC (darker), the more blurred the image of the specific target object OB becomes, so the accuracy of the distance ΔD decreases. In addition, the smaller the size of the specific target object OB in the foreground image PIC (the longitudinal dimension H of the rectangular region R), the more blurred the image of the specific target object OB becomes (the lower the resolution), so the accuracy of the distance ΔD decreases. Thus, the actual position, brightness BR, and longitudinal dimension H of the specific target object OB affect the detection accuracy of the distance ΔD, but the degree of influence in the first state is greater than that in the second state.

[0037] Accordingly, the ECU 10 sequentially obtains the determination result of whether the lower end line L1 of the rectangular region R overlaps with the pattern graphic image PTN from the target object recognition device 30. When the lower end line L1 overlaps with the pattern graphic image PTN, the ECU 10 corrects the average distance ΔDave based on the offset value OFS obtained from the map M1 (position deviation map) as described below.

[0038] As Figure 3 shown, the map M1 is a database that defines the relationship between the distance ΔD (distance level), brightness BR (brightness level), and longitudinal dimension H (longitudinal dimension level), and the offset value OFS. In the map M1, the distance ΔD is divided into multiple levels (for example, 5 levels (distance level 1 (short) to distance level 5 (long))). In addition, in the map M1, the brightness BR is divided into multiple levels (for example, 3 levels (brightness level 1 (bright) to brightness level 3 (dark))). In addition, the longitudinal dimension H is divided into multiple levels (for example, 6 levels (longitudinal dimension level 1 (large) to longitudinal dimension level 6 (small))). The map M1 is composed of a table that defines the relationship between each level of the distance ΔD, brightness BR, and longitudinal dimension H, and the offset value OFS. Specifically, the map M1 includes tables TD1 to TD5 corresponding to distance levels 1 to 5. These tables TD1 to TD5 are equivalent to tables divided by position. In addition, each table TDn (n = 1, 2,..., 5) is composed of tables TBR1, TBR2, and TBR3 corresponding to brightness levels 1 to 3. These tables TBR1 to TBR3 are equivalent to tables divided by brightness. Furthermore, each table TBRn (n = 1, 2, 3) is composed of longitudinal dimension tables THm (m = 1, 2,..., 6) that represent the relationship between longitudinal dimension levels 1 to 6 and the offset value OFS. These tables TH1 to TH6 are equivalent to tables divided by size.

[0039] Each offset value OFS is determined by performing a predetermined calibration process using a calibration device (computer) and six simulation bodies TS1 to TS6 of different heights (models of a specific target object OB) during the design phase of the own vehicle (or at the time of factory shipment). In addition, these six simulation bodies TS1 to TS6 correspond to longitudinal dimension levels 1 to 6. Hereinafter, the process (steps) of the calibration process will be described.

[0040] First, an experimental vehicle equipped with the vehicle control system 1 is installed in a predetermined laboratory. Next, a pattern graphic is projected onto the floor of the laboratory by the projection device 20. Next, the simulation body TS1 corresponding to the longitudinal dimension level 1 is arranged at a predetermined position within the pattern graphic projected onto the road surface (floor), that is, at the position P1 corresponding to the distance level 1. In this state, in the foreground image PIC, the lower end line L1 of the rectangular area R overlaps with the pattern graphic image PTN. Next, the illumination brightness of the laboratory is adjusted to be consistent with a predetermined brightness corresponding to the brightness level 1.

[0041] Next, the calibration device sequentially obtains the brightness BR of the foreground image PIC, the longitudinal dimension H of the rectangular area R, and the distance ΔD from the target recognition device 30. The calibration device calculates the average values of a predetermined number of distances ΔD, brightness BR, and longitudinal dimensions H obtained from the target recognition device 30, and stores the calculation results (average distance ΔD1ave, average brightness BRave, and average longitudinal dimension Have). Next, the operation of the projection device 20 is stopped so that the pattern graphic is not projected onto the road surface. In this state, the calibration device sequentially obtains the distances ΔD from the target recognition device 30, calculates their average value (average distance ΔD2ave), and stores the calculation result. Next, the calibration device obtains the deviation between the average distance ΔD1ave and the average distance ΔD2ave as the offset value OFS (refer to Figure 4 ). Then, the calibration device associates this offset value OFS with the average distance ΔD1ave, the average brightness BRave, and the average longitudinal dimension Have. By doing so, the offset value OFS for the distance level 1, the brightness level 1, and the longitudinal dimension level 1 is determined. In addition, this average distance ΔD1ave is the representative value d1 of the distance level 1. Further, this average brightness BRave is the representative value br1 of the brightness level 1. Further, this average longitudinal dimension Have is the representative value h1 of the longitudinal dimension level 1. The average distance ΔD1ave corresponds to the first position, and the average distance ΔD2ave corresponds to the second distance. In addition, the offset value OFS corresponds to the deviation amount between the first position and the second position.

[0042] Next, replacing the dummy body TS1, the dummy body TS2 corresponding to the longitudinal dimension level 2 is disposed at the position P1. In addition, the illumination brightness in the laboratory is maintained at the brightness corresponding to the brightness level 1. In this experimental environment, through the same process as the above process, the average distances ΔD1ave, ΔD2ave, etc. are obtained. Next, the calibration device obtains the deviation between the average distance ΔD1ave and the average distance ΔD2ave as the offset value OFS. Then, the calibration device associates this offset value OFS with the average distance ΔD1ave, the average brightness BRave, and the average longitudinal dimension Have. By doing so, the offset value OFS corresponding to the distance level 1, the brightness level 1, and the longitudinal dimension level 2 is determined. In addition, this average longitudinal dimension Have is the representative value h2 of the longitudinal dimension level 2.

[0043] Next, without changing the illumination brightness in the laboratory, the dummy bodies TS3, TS4, TS5, and TS6 are sequentially disposed at the position P1, and through the same process as the above process, the offset values OFS corresponding to the distance level 1, the distance level 1, and the longitudinal dimension levels m (m = 3, 4, 5, 6) are sequentially obtained.

[0044] Next, the illumination brightness in the laboratory is adjusted to the predetermined brightness corresponding to the brightness level 2. Moreover, the dummy bodies TS1 to TS6 are sequentially disposed at the position P1, and through the same process as the above process, the offset values OFS corresponding to the distance level 1, the brightness level 2, and the longitudinal dimension levels m (m = 1, 2,..., 6) are sequentially obtained. Next, the illumination brightness in the laboratory is adjusted to the predetermined brightness corresponding to the brightness level 3. Moreover, the dummy bodies TS1 to TS6 are sequentially disposed at the position P1, and through the same process as the above process, the offset values OFS corresponding to the distance level 1, the brightness level 3, and the longitudinal dimension levels m (m = 1, 2,..., 6) are sequentially obtained.

[0045] Next, the illumination brightness in the laboratory is readjusted to the predetermined brightness corresponding to the brightness level 1. Moreover, the dummy bodies TS1 to TS6 are sequentially disposed at the position P2 corresponding to the distance level 2, and through the same process as the above process, the offset values OFS corresponding to the distance level 2, the brightness level 1, and the longitudinal dimension levels m (m = 1, 2,..., 6) are sequentially obtained.

[0046] After that, the offset value OFS corresponding to other experimental environments (other combinations of distance levels, brightness levels, and longitudinal dimension levels) is obtained through the same process as the above process.

[0047] In each of the tables TBRa (a = 1, 2, 3) divided by brightness in the mapping M1 constructed through the above process, the offset value OFS of the longitudinal dimension level m (m = 1, 2, …) is smaller than the offset value OFS of the longitudinal dimension level n (m < n). In addition, the offset value OFS of the longitudinal dimension level m in the table TBRa divided by brightness in the distance table (table divided by position) TDx is equal to or less than the offset value OFS of the longitudinal dimension level m in the table TBRb (b > a) divided by brightness in this distance table TDx. In addition, the offset value OFS of the longitudinal dimension level m in the table TBRa divided by brightness in the distance table TDi (i = 1, 2, …) is equal to or less than the offset value OFS of the longitudinal dimension level m in the table TBRa divided by brightness in the distance table TDj (j > i). This mapping M1 is written into the ROM10b during the production of the vehicle V.

[0048] While the target object recognition device 30 is activated, the ECU10 sequentially obtains the distance ΔD from the target object recognition device 30. Furthermore, when the pattern graphic is projected onto the road surface by the projection device 20, the ECU10 sequentially obtains from the target object recognition device 30 the determination result regarding the overlap between the lower end line L1 of the rectangular area R in the foreground image PIC and the pattern graphic image PTN. When the ECU10 obtains the determination result indicating the overlap between the lower end line L1 and the pattern graphic image PTN, in addition to obtaining the distance ΔD, the ECU10 sequentially obtains from the target object recognition device 30 the brightness BR and the longitudinal dimension H. The ECU10 calculates the average distance ΔDave, the average brightness BRave, and the average longitudinal dimension Have, which are the averages of the distance ΔD, the brightness BR, and the longitudinal dimension H corresponding to a predetermined number of consecutive frames. The ECU10 determines the distance level among distance levels 1, 2, …, 5 whose representative value is closest to the current average distance ΔDave. The ECU10 determines the brightness level among brightness levels 1, 2, …, 3 whose representative value is closest to the current average brightness Brave. In addition, the ECU10 determines the longitudinal dimension level among longitudinal dimension levels 1, 2, …, 6 whose representative value is closest to the current value Have. The ECU10 obtains the offset value OFS corresponding to the determined (selected) distance level, brightness level, and longitudinal dimension level from the mapping M1. When the corrected average distance ΔDave obtained by subtracting the offset value OFS from the average distance ΔDave is equal to or less than the threshold value ΔDth, the ECU10 executes the second notification process.

[0049] On the other hand, the ECU 10 does not correct the average distance ΔDave without obtaining a determination result indicating that the lower end line L1 overlaps with the pattern graphic image PTN in the foreground image PIC. That is, the ECU 10 executes the second notification process when the average value of the distances ΔD sequentially obtained from the target recognition device 30, i.e., the average distance ΔD, is equal to or less than the threshold value ΔDth.

[0050] Next, Figure 5 a description will be given of a program PR1 executed by the CPU 10a (hereinafter simply referred to as "CPU") of the ECU 10 to implement the above-described correction function of the average distance ΔDave.

[0051] Program PR1

[0052] When the pattern graphic is projected onto the road surface by the projection device 20, the CPU starts the execution of the program PR1 at a predetermined cycle. The CPU starts the execution of the program PR1 from step 100 and advances the process to step 101.

[0053] In step 101, the CPU obtains from the target recognition device 30 a determination result regarding the overlap of the lower end line L1 of the rectangular region R and the pattern graphic image PTN. When the CPU obtains a determination result indicating that the lower end line L1 overlaps with the pattern graphic image PTN (101: Yes), the process advances to step 102. On the other hand, when the CPU does not obtain a determination result indicating that the lower end line L1 overlaps with the pattern graphic image PTN (101: No), the process advances to step 105.

[0054] In step 102, the CPU obtains the distance ΔD, the brightness BR, and the longitudinal dimension H from the target recognition device 30. Each time the CPU obtains this information, the CPU stores this information in the RAM 10c (circular buffer). As a result, the distance ΔD, the brightness BR, and the longitudinal dimension H (time series data) corresponding to a predetermined number of frames (foreground image PIC) are stored in the RAM 10c. The CPU calculates the average distance ΔDave, the average brightness BRave, and the average longitudinal dimension Have based on this time series data. Next, the CPU advances the process to step 103.

[0055] In step 103, the CPU obtains the offset value OFS with reference to the map M1. That is, the CPU respectively determines the distance level, the brightness level, and the longitudinal dimension level of the average distance ΔDave, the average brightness BRave, and the average longitudinal dimension Have. Then, the CPU obtains the offset value OFS corresponding to each determined level from the map M1. Next, the CPU advances the process to step 104.

[0056] The CPU uses the value obtained by subtracting the offset value OFS from the average distance ΔDave in step 104 as the corrected average distance ΔDave. Then, the CPU advances the process to step 106, in which the execution of program PR1 (the correction process of the average distance ΔDave) is terminated.

[0057] When the CPU advances the process from step 101 to step 105, it acquires the distance ΔD from the target recognition device 30 and stores it in the RAM 10c. Further, the CPU calculates the average distance ΔDave based on the time-series data of the distance ΔD stored in the RAM 10c. Then, without correcting the average distance ΔDave, the CPU advances the process to step 106, in which the execution of program PR1 is terminated.

[0058] Effect

[0059] As described above, the vehicle control system 1 includes the projection device 20 that projects a predetermined pattern on the road surface in front of the host vehicle, and the target recognition device 30 that acquires the distance ΔD as the position information indicating the relative position between the host vehicle and the specific target object OB based on the foreground image PIC. Here, in the foreground image PIC, when the lower end line L1 of the rectangular area R overlaps with the pattern graphic image PTN, the accuracy of the position information may sometimes be low. According to the vehicle control system 1, when the two images overlap, the average distance ΔDave is corrected by a predetermined correction process. Thereby, a decrease in the detection accuracy of the distance ΔD between the specific target object OB and the host vehicle is suppressed.

[0060] The present invention is not limited to the above-described embodiment, and various modifications can be adopted within the scope of the present invention as described below.

[0061] Modification Example 1

[0062] In the above-described embodiment, in the mapping M1, the distance ΔD is classified into distance levels 1 to 5, but the number of distance levels (the number of divisions) can also be changed. For example, in the mapping M1, the distance ΔD can also be classified into more levels than in the above-described embodiment. Further, in the mapping M1, the number of brightness levels and the number of vertical size levels can also be changed.

[0063] Modification Example 2

[0064] In the above-described embodiment, when a predetermined condition is satisfied, the ECU 10 causes the projection device 20 to project a graphic indicating the traveling direction of the host vehicle onto the road surface. Instead, when the predetermined condition is satisfied, the ECU 10 may also cause the projection device 20 to project a pattern graphic for detecting road surface unevenness onto the road surface. In this case, the target recognition device 30 detects the road surface unevenness based on the deformation of the pattern graphic image PTN in the foreground image PIC.

[0065] Modification Example 3

[0066] In the above-described embodiment, when the average distance ΔD is equal to or less than the threshold value ΔDth, the second notification process is executed. However, instead of this, or in addition to this, an automatic braking process for automatically braking the host vehicle may be executed.

[0067] Modification Example 4

[0068] In the above-described embodiment, the projection device 20 faces the front of the host vehicle, and the imaging device of the target recognition device 30 also faces the front of the host vehicle. Instead, these devices may also face the rear of the host vehicle. That is, the projection device 20 may project a predetermined pattern graphic onto the road surface behind the host vehicle. Moreover, the target recognition device 30 may be able to recognize a specific target located behind the host vehicle.

Claims

1. A vehicle control system comprising: a projection device that projects a predetermined pattern onto the road surface around the host vehicle; a target object recognition device that recognizes a specific target object located around the host vehicle based on a surrounding image obtained by photographing the surrounding area of the host vehicle, and outputs information on the relative position of the host vehicle and the specific target object, i.e., position information; and a processor that controls the projection device and the target object recognition device, wherein the processor is configured to perform a predetermined correction process including correcting the relative position obtained based on the position information when a graphic image projected by the projection device onto the road surface overlaps an object recognition image, which is an image of an area recognized as the specific target object in the surrounding image.

2. The vehicle control system according to claim 1, comprising a position deviation map that defines a deviation amount between a first position and a second position, the first position and the second position being respectively obtained based on the position information output from the target object recognition device in the same position relationship between the host vehicle and the specific target object, the first position being the position of the specific target object relative to the host vehicle obtained based on the position information output when the graphic image overlaps the object recognition image in a first state, and the second position being the position of the specific target object relative to the host vehicle obtained based on the position information output when the graphic image does not overlap the object recognition image in a second state, The correction process includes: and correcting the first position obtained in the first state based on the position deviation map.

3. The vehicle control system according to claim 2, wherein the position deviation map includes a plurality of position deviation tables selected based on at least one of the image size of the object recognition image, the brightness of the surrounding image, and the first position.

4. The vehicle control system according to claim 2, wherein the processor obtains the first position, the brightness of the surrounding image, and the image size of the object recognition image, the position deviation map includes a plurality of position-divided tables selected based on the first position, each position-divided table includes a plurality of brightness-divided tables selected based on the brightness of the surrounding image, and each brightness-divided table includes a plurality of size-divided tables that define the deviation amount selected based on the image size.

5. The vehicle control system according to any one of claims 1 to 4, wherein the object recognition image is a rectangular image, and the processor performs the correction process when the graphic image overlaps the lower end line of the object recognition image.

Images