JP4882324B2 - Image processing apparatus for vehicle - Google Patents

Description

  The present invention relates to a vehicle image processing apparatus that optimally supports vehicle control based on, for example, a stereo image captured by an in-vehicle camera.

2. Description of the Related Art Conventionally, a distance detection device that can correct an error caused by a deviation in parallelism between optical axes of two imaging devices to calculate an accurate parallax and improve distance detection accuracy is known (for example, Patent Documents). 1). The inter-vehicle distance measuring device and the distance detecting device calculate the distance to the object based on the stereo image.
JP 2001-169310 A

  However, in the conventional technique shown in Patent Document 1, when there are a plurality of objects having the same shape around the vehicle, a plurality of corresponding point candidates in the stereo image may be detected. In this case, it is conceivable that all candidates are not adopted and detection of corresponding points is repeated, but detection of corresponding points becomes impossible, and there is a possibility that the distance to the target cannot be calculated. On the other hand, it is conceivable to calculate the degree of correlation (similarity) for a plurality of corresponding point candidates and adopt the one with the larger degree of correlation, but the degree of correlation is high due to the influence of noise such as ambient light. Even if one is selected, a true corresponding point cannot always be selected. For this reason, there is a possibility that the corresponding point is erroneously detected, and the distance from the object is erroneously calculated. Therefore, when the calculated distance to the object is used for vehicle control, the vehicle control may not be optimally performed.

  The present invention has been made to solve such problems, and has as its main purpose to optimally support vehicle control.

In order to achieve the above object, one embodiment of the present invention provides:
Stereo image acquisition means for acquiring a stereo image around the host vehicle;
Corresponding point detecting means for detecting corresponding points between the stereo images acquired by the stereo image acquiring means;
A distance calculation unit that calculates a parallax based on the corresponding point detected by the corresponding point detection unit, and calculates a distance between the host vehicle and an object around the host vehicle based on the parallax;
A vehicle image processing apparatus comprising:
Corresponding point selection that selects one corresponding point from the plurality of corresponding point candidates based on the control mode of the host vehicle when the corresponding point detecting means detects a plurality of corresponding point candidates. An image processing apparatus for a vehicle characterized by comprising means.

  According to this aspect, when the corresponding point detecting unit detects a plurality of corresponding point candidates, the corresponding point selecting unit selects one of the plurality of corresponding point candidates based on the control mode of the host vehicle. Select the corresponding point. Thereby, an optimal corresponding point is selected according to control modes, such as PCS control, ACC control, parking assistance control, etc., for example. Therefore, vehicle control can be optimally supported.

  In this aspect, the control mode of the host vehicle is a pre-crash (PCS) control that adds deceleration to the host vehicle and / or activates an occupant protection device when a collision with an object is estimated. In this case, it is preferable that the corresponding point selection unit selects the corresponding point that is far from the host vehicle among the plurality of corresponding point candidates. As a result, when the corresponding point selected by the corresponding point selection means matches the true corresponding point, the collision is properly determined based on the actual distance from the object, and normal PCS control is executed. can do. On the other hand, for example, even if the corresponding point selected by the corresponding point selection means is not a true corresponding point, the corresponding point that is far from the host vehicle is intentionally selected, so the distance to the object is actually Recognized farther. Therefore, in the operation of the occupant protection device, it is possible to reliably prevent control delays and early operations such as pre-crash control that are likely to cause the passengers to feel discomfort due to delays in operation or early operations. can do. That is, the vehicle control can be optimally supported.

  Furthermore, in this one aspect, when the control mode of the host vehicle is an inter-vehicle control (ACC control) that appropriately maintains a distance from the preceding vehicle, the corresponding point selection unit is configured to select the plurality of corresponding points. Of the candidates, it is preferable to select the corresponding point that is close to the host vehicle. As a result, when the corresponding point selected by the corresponding point selecting means matches the true corresponding point, the normal ACC control can be executed based on the actual distance from the object such as the preceding vehicle. it can. On the other hand, for example, even if the corresponding point selected by the corresponding point selection means is not a true corresponding point, the corresponding point that is close to the host vehicle is selected, so that the distance to the target is recognized closer than the actual distance. Is done. Therefore, ACC control is executed so that a sufficient distance from the object is ensured, leading to an improvement in safety. That is, the vehicle control can be optimally supported.

  In this aspect, when the control mode of the host vehicle is parking support control that guides the host vehicle to a predetermined parking position, the corresponding point selection unit includes the host vehicle among the plurality of corresponding point candidates. It is preferable to select the corresponding points that are close to each other. Thereby, when the corresponding point selected by the corresponding point selecting means matches the true corresponding point, based on the actual distance from the object such as a person or an obstacle on the route to the parking position, Normal parking assistance control can be executed. On the other hand, for example, even if the corresponding point selected by the corresponding point selection means is not a true corresponding point, the corresponding point that is close to the host vehicle is selected, so that the distance to the target is recognized closer than the actual distance. Is done. Therefore, parking support control is executed so that a safe distance from the object is sufficiently secured. That is, the vehicle control can be optimally supported.

  In this aspect, the corresponding point selection unit may select one corresponding point from the plurality of corresponding point candidates in accordance with the traveling environment of the host vehicle. The traveling environment of the host vehicle refers to, for example, brightness around the host vehicle, ease of slipping on the road surface on which the host vehicle is traveling, vehicle speed, vehicle acceleration, and the like. Further, for example, the corresponding point selection means may be selected from a plurality of corresponding point candidates that the surroundings of the host vehicle become dark, the road surface becomes slippery, the vehicle speed increases, and / or the vehicle acceleration increases. A corresponding point that is far from the host vehicle may be selected.

  In this aspect, the corresponding point at a position far from the host vehicle indicates a corresponding point with a small parallax, and the corresponding point at a position near the host vehicle indicates a corresponding point with a large parallax.

  According to the present invention, vehicle control can be optimally supported.

  Hereinafter, the best mode for carrying out the present invention will be described with reference to the accompanying drawings. Note that the basic concept of the image processing apparatus, the main hardware configuration, the operating principle, the basic control method, and the like are known to those skilled in the art, and thus detailed description thereof is omitted.

  FIG. 1 is a block diagram showing an example of the system configuration of a vehicle image processing apparatus 1 according to the first to third embodiments of the present invention. The vehicular image processing apparatus 1 according to the first to third embodiments includes photographing means such as two cameras 3R and 3L that photograph the surroundings of the host vehicle.

  The cameras 3R and 3L perform image processing on the images taken by the cameras 3R and 3L, and objects around the vehicle such as preceding vehicles, people, obstacles, white lines (hereinafter referred to as objects). An image processing ECU (Electronic Control Unit) 5 that calculates the distance between the vehicle and the host vehicle is connected.

  The cameras 3R and 3L are arranged on the right and left sides of the front part of the vehicle, for example, so as to face the front of the vehicle, the lens optical axes of the cameras 3R and 3L are parallel to each other, and the horizontal axis or vertical axis of the imaging surface is It arrange | positions so that it may align on the same line (FIG. 2).

  Note that although two cameras 3R and 3L are used as photographing means, a single stereo camera including a stereo optical system having two lenses may be used. The stereo camera is advantageous in saving space because a stereo image is acquired by a single camera having a plurality of viewpoints.

  Further, as the cameras 3R and 3L, a CCD (Charge Coupled Device) or CMOS (Complementary Metal Oxide Semiconductor) camera that detects visible light is used, but an infrared camera that is effective for shooting at night may be used. Furthermore, although the cameras 3R and 3L are disposed at the front part of the vehicle, they may be disposed at any position of the vehicle such as the rear part of the vehicle or the side part of the vehicle. Stereo image information photographed by each of the cameras 3R and 3L is transmitted to the image processing ECU 5.

  The image processing ECU 5, a collision-response ECU 7, which will be described later, a seat belt ECU 9, an air bag ECU 11, a brake ECU 13, an engine ECU 15, a transmission ECU 17, an inter-vehicle ECU 19, a parking assistance ECU 21, and a steering ECU 23 are composed of a microcomputer and A CPU for controlling each part of the device, a ROM for storing an execution program of the CPU, a readable / writable RAM (Random Access Memory) for storing image data, calculation results, a timer, a counter, an input It has an output interface. The CPU, ROM, RAM, and input / output interface are connected to each other by a data bus.

  The input / output interface includes an A / D (Analog / Digital) converter that converts an analog signal into a digital signal, a D / A (Digital / Analog) converter that converts a digital signal into an analog signal, and the like. Analog image data transmitted from the cameras 3R and 3L is converted into digital image data by the A / D converter of the input / output interface, and stored in the RAM image memory. The CPU reads out the image data stored in the image memory of the RAM and performs image processing according to the image processing program stored in the ROM. Note that the image data stored in the image memory is a grayscale image including luminance information.

  The image processing ECU 5 generates a binarized image based on the gray scale image read from the image memory. Specifically, in the grayscale image, the image processing ECU 5 sets a region (pixel) larger than a preset luminance threshold to “1” and sets a small region (pixel) to “0” to convert the binarized image into a grayscale image. Generate. The image processing ECU 5 performs various image processing described later on these gray scale images or binarized images.

  Note that the image processing ECU 5, a collision response ECU 7, which will be described later, a seat belt ECU 9, an airbag ECU 11, a brake ECU 13, an engine ECU 15, a transmission ECU 17, an inter-vehicle ECU 19, a parking assist ECU 21, and a steering ECU 23 are connected to a vehicle LAN (Local Area Network) 25. For example, two-way data communication is performed based on a CAN (Controller Area Network) protocol.

  Next, a method for calculating the distance S between the host vehicle and the object by the image processing ECU 5 configured as described above will be described.

  First, the image processing ECU 5 is a shift in the left and right direction (horizontal direction) on the same object with respect to the same object in the right image taken by the right camera 3R and the left image taken by the left camera 3L. The parallax Δd is calculated. The image processing ECU 5 calculates the parallax Δd for each predetermined region on the image (for example, for each of a plurality of pixels).

  For example, the image processing ECU 5 uses the left camera 3L for an object A such as a preceding vehicle extracted on an image (hereinafter referred to as a reference image) captured by the right camera 3R. A corresponding point (hereinafter referred to as a corresponding image) B on the search image) is detected. Note that an image captured by the left camera 3L may be used as a reference image, and an image captured by the right camera 3R may be used as a search image.

  Specifically, the image processing ECU 5 detects the corresponding image B by searching an area on the search image that is most similar to the object A of the reference image. Note that the degree of similarity (correlation) between images is calculated using a difference process based on the luminance value of each pixel in the region, a normalized correlation process, or the like. For example, the image processing ECU 5 calculates a sum of luminance differences (luminance difference sum value) for each pixel in the region in the difference processing. The similarity increases as the calculated luminance difference total value decreases. Note that the method of calculating the luminance difference total value is a well-known technique, and thus detailed description thereof is omitted.

  The image processing ECU 5 is based on the distance ax between the centroid position of the object A in the reference image and the image center line L, and the distance bx between the centroid position of the object image B in the search image and the image center line L. The parallax Δd is calculated by the following equation (1) (FIG. 3).

Δd = ax + bx (1)
In this way, the image processing ECU 5 performs image processing on the same object in the reference image (right image) captured by the right camera 3R and the search image (left image) captured by the left camera 3L. The parallax Δd, which is a deviation in the horizontal direction (horizontal direction), is calculated.

  Further, the image processing ECU 5 uses the calculated parallax Δd and the following equation (2) based on the principle of triangulation to calculate the distance between the object and the host vehicle (hereinafter referred to as an inter-vehicle distance). ) Calculate S.

S = f × D / (Δd × p) (2)
In the above equation (2), f is the focal length of the left and right cameras 3R and 3L, and D is the distance between the optical axes of the left and right cameras 3R and 3L. Further, p is a pixel interval in the image sensor 3a (FIG. 4). Note that values obtained experimentally in advance are set in the ROM of the image processing ECU 5 as the focal length f, the distance D between the optical axes, and the pixel interval p.
(First embodiment)
The image processing ECU 5 adds a deceleration to the own vehicle based on the inter-vehicle distance S calculated by the image processing ECU 5 and the relative speed Vr of the own vehicle with respect to an object such as a preceding vehicle, and an occupant protection device is installed. A collision-response ECU 7 that performs PCS (Pre Crush Safety) control such as operation is connected. The image processing ECU 5 calculates a relative speed Vr of the host vehicle with respect to the object by performing a differentiation process on the inter-vehicle distance S.

  More specifically, the collision handling ECU 7 predicts a collision between the vehicle front, rear, or side object and the host vehicle based on the inter-vehicle distance S calculated by the image processing ECU 5 and the relative speed Vr. If this happens, add deceleration to the host vehicle and / or activate and control the occupant protection device. Note that adding deceleration to the host vehicle includes preparation for adding deceleration, and operating the occupant protection device includes preparation for operation.

  The occupant protection device includes, for example, a seat belt device 9a for restraining the occupant to the vehicle seat, an airbag device 11a for expanding and inflating the airbag to absorb the impact on the occupant, and an impact absorbing mechanism for the steering column. All devices for ensuring vehicle safety, such as a built-in steering device 23a, a brake pedal device that retracts and absorbs the impact when an impact occurs, and a brake lamp that informs the rear vehicle of deceleration (depression of the brake pedal) of the host vehicle Is included.

  Note that the collision handling ECU 7 starts PCS control such as operating the occupant protection device at an earlier timing as the inter-vehicle distance S is shorter and / or as the relative speed Vr is faster, for example. On the contrary, the collision handling ECU 7 starts PCS control such as operating the occupant protection device at a later timing as the inter-vehicle distance S is longer and / or as the relative speed Vr is slower.

  In other words, if the inter-vehicle distance S calculated by the image processing ECU 5 is calculated to be shorter than the actual inter-vehicle distance, the PCS control such as actuating the occupant protection device is started intentionally at an earlier timing. be able to. Conversely, if the inter-vehicle distance S calculated by the image processing ECU 5 is calculated to be longer than the actual inter-vehicle distance, PCS control such as intentionally operating the occupant protection device at a later timing is started. Can do.

  A seatbelt ECU 9 for operating and controlling the seatbelt device 9a is connected to the collision handling ECU7. The seat belt device 9a incorporates a winding device (pretensioner) that applies tension to the seat belt. When it is predicted that the subject and the host vehicle will collide based on the inter-vehicle distance S from the image processing ECU 5 and the relative speed Vr, the collision handling ECU 7 applies to the seat belt ECU 9, an air bag ECU 11 and a brake ECU 13 described later. Send a collision signal. When the seat belt ECU 9 receives a collision signal from the collision handling ECU 7, the seat belt ECU 9 operates the pretensioner of the seat belt device 9a to apply tension to the seat belt. This ensures that the occupant is restrained by the seat belt in the event of a vehicle collision.

  An air bag ECU 11 that operates and controls the air bag device 11a is connected to the collision handling ECU 7. When the airbag ECU 11 receives the collision signal from the collision handling ECU 7, the airbag ECU 11 prepares for the operation of the airbag apparatus 11a. Thus, the airbag apparatus 11a can be reliably operated before and after the host vehicle collides with the object.

  A brake ECU 13 for controlling a hydraulic brake device 13a of the host vehicle is connected to the collision-responsive ECU 7. The brake ECU 13 controls the deceleration of the vehicle by, for example, controlling the hydraulic pressure of the brake device 13a and operating the brake caliper. When the brake ECU 13 receives a collision signal from the collision handling ECU 7, the brake ECU 13 controls the hydraulic pressure of the brake device 13a to add deceleration to the host vehicle. Thereby, the impact when the own vehicle collides with the object can be reduced.

  The collision handling ECU 7 applies deceleration to the host vehicle by controlling the brake device 13a via the brake ECU 13, but operates a regenerative brake by power generation of the motor that reduces the output of the engine 15a ( For example, in the case of an electric vehicle, a hybrid vehicle that travels by using an engine and a motor together), a deceleration can be added to the host vehicle by changing the transmission ratio of the transmission device 17a or any combination thereof. Good.

  Specifically, an engine ECU 15 that controls the engine 15a is connected to the collision-compatible ECU 7. When the engine ECU 15 receives a collision signal from the collision-compatible ECU 7, the engine-controlled throttle actuator is controlled to output the engine 15a. It may be lowered. In addition, a transmission ECU 17 for controlling the transmission 17a of the vehicle is connected to the collision-compatible ECU 7. When the transmission ECU 17 receives the collision signal, the transmission ECU 17 changes the gear ratio of the transmission mechanism to add deceleration to the vehicle. May be.

  By the way, the image processing ECU 5 detects a corresponding image having a high similarity (for example, a small luminance difference sum value, for example, a value close to 0) on the search image with respect to the object on the reference image. In this case, there may be a plurality of corresponding image candidates with high similarity (small luminance difference sum value) on the search image. In addition, when the difference of the luminance difference sum value is small between the corresponding image candidates (for example, within the range of the error), and when noise is further mixed (for example, the influence of the disturbance light, the difference between the left and right imaging characteristics), In the related art, there is a possibility that an appropriate corresponding image is not selected.

  Therefore, in this embodiment, the image processing ECU 5 selects one corresponding image from among a plurality of corresponding image candidates based on the control mode of the host vehicle such as PCS control.

  Specifically, the image processing ECU 5 calculates the inter-vehicle distance S for each corresponding image candidate based on the above equation (1). Further, the image processing ECU 5 selects a corresponding image having a large inter-vehicle distance S and a distance from the own vehicle (for example, a corresponding image having the maximum inter-vehicle distance S) from the corresponding image candidates. The inter-vehicle distance S of the corresponding image is transmitted to the collision corresponding ECU 7 that performs PCS control. The collision handling ECU 7 performs the above-described PCS control based on the transmitted inter-vehicle distance S far from the host vehicle.

  In addition, although the example which selects the corresponding image in which the inter-vehicle distance S is the maximum as the corresponding image far from the host vehicle is given, the average value of the inter-vehicle distances S among a plurality of corresponding image candidates is calculated, and this average is calculated. A corresponding image having an inter-vehicle distance S larger than the value may be arbitrarily selected.

  Thereby, when the corresponding image selected by the image processing ECU 5 matches the actual object, the collision handling ECU 7 appropriately determines the collision based on the inter-vehicle distance S that is the actual inter-vehicle distance, Perform normal PCS control. On the other hand, for example, even when the corresponding image selected by the image processing ECU 5 is different from the actual target object and the inter-vehicle distance S is different from the actual inter-vehicle distance, the collision corresponding ECU 7 is intentionally more than the actual inter-vehicle distance. Based on the large inter-vehicle distance S, a collision is determined and PCS control is performed. Therefore, it is determined that there is a possibility that the collision handling ECU 7 may collide based on the inter-vehicle distance S shorter than the actual inter-vehicle distance, and there is a sense of incongruity due to the general activation of the occupant protection device or the like. It is possible to prevent the PCS control that is easily performed from being operated unnecessarily early.

  Next, a processing flow of the vehicle image processing apparatus 1 according to the first embodiment will be described. FIG. 5 is a flowchart showing an example of a processing flow of the vehicle image processing apparatus 1 according to the first embodiment. Note that the processing routines shown in FIG. 5, FIG. 6 and FIG. 7 to be described later are repeatedly executed every predetermined minute time, for example, every 64 ms.

  The image processing ECU 5 acquires a stereo image around the vehicle from each camera 3R, 3L (S100).

  Next, the image processing ECU 5 detects a corresponding image on the search image for an object such as a preceding vehicle extracted on the reference image based on the acquired stereo image (S110).

  Thereafter, the image processing ECU 5 determines whether or not a plurality of corresponding image candidates are detected on the search image (S120).

  When the image processing ECU 5 determines that a plurality of corresponding image candidates are detected (Yes in S120), the image processing ECU 5 calculates the inter-vehicle distance S for each corresponding image based on the above (1) (S130). On the other hand, when the image processing ECU 5 determines that only one corresponding image candidate has been detected (No in S120), the inter-vehicle distance S of the corresponding image is calculated based on the above equation (1), and will be described later (S150). )

  Next, the image processing ECU 5 selects a correspondence image that maximizes the inter-vehicle distance S from among a plurality of correspondence image candidates (S140), and sets the inter-vehicle distance S of the selected correspondence image to the collision correspondence ECU 7. Send.

  The collision handling ECU 7 determines whether or not the host vehicle collides with the object based on the inter-vehicle distance S and the relative speed Vr transmitted from the image processing ECU 5 (S150).

  When it is determined that the host vehicle collides with the object (Yes in S150), the collision handling ECU 7 transmits a collision signal to the seat belt ECU 9, the airbag ECU 11, the brake ECU 13, and the like to execute PCS control (S160). For example, the seat belt ECU 9 operates the seat belt device 9a based on the collision signal. Further, the airbag ECU 11 prepares to operate the airbag device 11a based on the collision signal. The brake ECU 13 activates the brake device 13a based on the collision signal, and adds deceleration to the vehicle.

  On the other hand, when the collision handling ECU 7 determines that the host vehicle does not collide with the object (No in S150), the collision handling ECU 7 ends the processing routine.

As described above, in the vehicular image processing apparatus according to the first embodiment, the image processing ECU 5 selects and selects the farther corresponding image (corresponding image having a small parallax Δd) among the plurality of corresponding image candidates. The corresponding inter-vehicle distance S of the corresponding image is transmitted to the collision-compatible ECU 7. The collision handling ECU 7 executes PCS control based on the inter-vehicle distance S of the farther correspondence image transmitted. Thereby, when the corresponding image selected by the image processing ECU 5 matches the actual object, the collision handling ECU 7 appropriately determines the collision based on the inter-vehicle distance S that is the actual inter-vehicle distance, Perform normal PCS control. On the other hand, for example, even when the corresponding image selected by the image processing ECU 5 is different from the actual target object and the inter-vehicle distance S is different from the actual inter-vehicle distance, the collision corresponding ECU 7 is intentionally more than the actual inter-vehicle distance. Based on the large inter-vehicle distance S, a collision is determined and PCS control is executed. Therefore, it is determined that there is a possibility that the collision handling ECU 7 may collide based on the inter-vehicle distance S shorter than the actual inter-vehicle distance, and there is a sense of incongruity due to the general activation of the occupant protection device or the like. It is possible to prevent the PCS control that is easily performed from being operated unnecessarily early. That is, the vehicle control can be optimally supported.
(Second Embodiment)
When the image processing ECU 5 detects a plurality of corresponding image candidates, the image processing ECU 5 selects a corresponding image closer to the host vehicle from the corresponding image candidates (for example, a corresponding image with the minimum inter-vehicle distance S), and selects the selected image. The inter-vehicle distance S of the corresponding image thus transmitted is transmitted to the inter-vehicle ECU 19 that performs ACC control.

  In addition, although the example which selects the corresponding image in which the inter-vehicle distance S is the minimum as the corresponding image closer to the host vehicle has been given, the average value of the inter-vehicle distances S in a plurality of corresponding image candidates is calculated, and this average A corresponding image having an inter-vehicle distance S smaller than the value may be arbitrarily selected.

  The inter-vehicle ECU 19 is based on the inter-vehicle distance S transmitted from the image processing ECU 5 and the vehicle speed of the host vehicle detected by the vehicle speed sensor 15b (the inter-vehicle distance, the inter-vehicle time, ACC (Adaptive Cruise Control or Auto Cruise Control) control is performed to maintain the vehicle head time etc. properly.

  Specifically, the inter-vehicle ECU 19 determines the appropriate inter-vehicle distance (for example, a map, a table, etc.) based on the vehicle speed of the host vehicle detected by the vehicle speed sensor 15b and the relationship between the inter-vehicle distance and the vehicle speed set in advance (for example, a map, a table, etc.). (Optimal distance considering vehicle safety, fuel consumption, etc.) is calculated. Further, the inter-vehicle ECU 19 adds acceleration or deceleration to the vehicle so that the inter-vehicle distance S transmitted from the image processing ECU 5 becomes the calculated appropriate inter-vehicle distance.

  For example, the inter-vehicle ECU 19 transmits a control signal to the engine ECU 15, the brake ECU 13, the transmission ECU 17, and the like to control the engine 15a, the brake device 13a, and the transmission device 17a, thereby adding deceleration to the vehicle. On the other hand, the inter-vehicle ECU 19 transmits a control signal to the engine ECU 15 to increase the output of the engine 15a, thereby adding acceleration to the vehicle.

  The inter-vehicle time is a value calculated by dividing the inter-vehicle distance by the vehicle speed of the host vehicle detected by the vehicle speed sensor 15b. The vehicle head time is a vehicle speed sensor 15b that detects the inter-vehicle distance between the host vehicle and the preceding vehicle that is normally maintained by the driver when the host vehicle is following the vehicle at (substantially) the same speed as the preceding vehicle. The value is divided by the speed of the subject vehicle.

  Other configurations are substantially the same as those of the first embodiment. The same parts as those in the first embodiment are denoted by the same reference numerals, and detailed description thereof is omitted.

  FIG. 6 is a flowchart illustrating an example of a processing flow of the vehicle image processing apparatus 1 according to the second embodiment. The processes (S200) to (S230) according to the second embodiment shown in FIG. 6 are the same as the processes (S100) to (S130) according to the first embodiment shown in FIG. Detailed description will be omitted.

  When the image processing ECU 5 determines that a plurality of corresponding image candidates are detected (Yes in S220), the image processing ECU 5 calculates the inter-vehicle distance S for each corresponding image based on (1) above (S230). On the other hand, if the image processing ECU 5 determines that only one corresponding image candidate has been detected (No in S220), the inter-vehicle distance S of the corresponding image is calculated based on the above equation (1), and (S250). Transition to processing.

  Next, the image processing ECU 5 selects a corresponding image that minimizes the inter-vehicle distance S from among a plurality of corresponding image candidates (S240), and transmits the inter-vehicle distance S of the selected corresponding image to the inter-vehicle ECU 19. To do.

  The inter-vehicle ECU 19 transmits a control signal to the engine ECU 15, the brake ECU 13, the transmission ECU 17, etc. so that the inter-vehicle distance S transmitted from the image processing ECU 5 becomes the calculated appropriate inter-vehicle distance, and the acceleration or reduction of the vehicle The speed is added and the ACC control is executed (S250). In this way, an appropriate inter-vehicle distance is maintained between an object such as a preceding vehicle and the host vehicle.

  Next, a modification of the vehicle image processing device 1 according to the second embodiment will be described.

  When the image processing ECU 5 detects a plurality of corresponding image candidates, the image processing ECU 5 may select a corresponding image closer to the own vehicle from among the corresponding image candidates according to the traveling environment of the vehicle.

  For example, when the image processing ECU 5 detects a plurality of corresponding image candidates, according to the vehicle speed from the vehicle speed sensor 15b or the vehicle acceleration from the acceleration sensor, the closer one of the corresponding image candidates from the own vehicle. An image may be selected. Thereby, a more appropriate corresponding image is selected according to the vehicle speed and the vehicle acceleration.

  As described above, in the vehicular image processing apparatus according to the second embodiment, the image processing ECU 5 selects the closest corresponding image (corresponding image having the larger parallax Δd) among the plurality of corresponding image candidates, The inter-vehicle distance S of the selected corresponding image is transmitted to the inter-vehicle ECU 19. The inter-vehicle ECU 19 executes ACC control based on the inter-vehicle distance S of the transmitted corresponding image.

As a result, when the corresponding image selected by the image processing ECU 5 matches the actual object, the inter-vehicle ECU 19 determines the inter-vehicle distance from the preceding vehicle or the like based on the inter-vehicle distance S that is the actual inter-vehicle distance. Judging properly, normal ACC control can be executed. On the other hand, for example, even when the corresponding image selected by the image processing ECU 5 is different from the actual object and the inter-vehicle distance S is different from the actual inter-vehicle distance, the inter-vehicle ECU 19 is intentionally smaller than the actual inter-vehicle distance. Based on the inter-vehicle distance S, the inter-vehicle distance from the preceding vehicle or the like is determined, and ACC control is executed. In this case, since the inter-vehicle ECU 19 recognizes the inter-vehicle distance S that is smaller than the actual inter-vehicle distance, the inter-vehicle distance control is performed so that the inter-vehicle distance with the preceding vehicle or the like becomes larger than usual. Therefore, since the sufficient distance between the vehicle and the preceding vehicle can be ensured, the safety of the vehicle is improved. That is, the vehicle control can be optimally supported.
(Third embodiment)
When the image processing ECU 5 detects a plurality of corresponding image candidates, the image processing ECU 5 selects a corresponding image closer to the host vehicle from the corresponding image candidates (for example, a corresponding image with the minimum inter-vehicle distance S), and selects the selected image. The inter-vehicle distance S of the corresponding image thus transmitted is transmitted to the parking support ECU 21 that performs parking support.

  In addition, although the example which selects the corresponding image in which the inter-vehicle distance S is the minimum as the corresponding image closer to the host vehicle has been given, the average value of the inter-vehicle distances S in a plurality of corresponding image candidates is calculated, and this average A corresponding image having an inter-vehicle distance S smaller than the value may be arbitrarily selected.

  The parking assist ECU 21 controls the steering ECU 23, the brake ECU 13, and the engine ECU 15 based on the inter-vehicle distance S between the host vehicle and the obstacle transmitted from the image processing ECU 5, and sets the parking target position set by the user. Then guide your vehicle. The steering ECU 23 controls the steering angle of the steering of the vehicle by controlling the steering device 23a.

  Other configurations are substantially the same as those of the first and second embodiments. The same parts as those in the first and second embodiments are denoted by the same reference numerals, and detailed description thereof is omitted.

  FIG. 7 is a flowchart showing an example of a processing flow of the vehicle image processing apparatus 1 according to the third embodiment. Note that the processes (S300) to (S330) according to the third embodiment shown in FIG. 7 are the same as the processes (S100) to (S130) according to the first embodiment shown in FIG. Therefore, detailed description is omitted.

  If the image processing ECU 5 determines that a plurality of corresponding image candidates have been detected (Yes in S320), the image processing ECU 5 calculates an inter-vehicle distance S for each corresponding image based on (1) above (S330). On the other hand, when the image processing ECU 5 determines that only one corresponding image candidate has been detected (No in S320), the inter-vehicle distance S of the corresponding image is calculated based on the above equation (1), and (S350). Transition to processing.

  Next, the image processing ECU 5 selects a corresponding image that minimizes the inter-vehicle distance S from among a plurality of corresponding image candidates (S340), and sets the inter-vehicle distance S of the selected corresponding image to the parking assist ECU 21. Send.

  The parking assist ECU 21 controls the steering ECU 23, the brake ECU 13, and the engine ECU 15 based on the inter-vehicle distance S transmitted from the image processing ECU 5, and executes parking assist control for guiding the host vehicle to the parking target position (S350). . Specifically, the parking assist ECU 21 avoids while maintaining a safe distance from obstacles such as people and other vehicles on the route to the parking target position based on the inter-vehicle distance S from the image processing ECU 5. Next, the steering ECU 23, the brake ECU 13, and the engine ECU 15 are controlled.

  As described above, in the vehicular image processing apparatus 1 according to the third embodiment, when the corresponding image selected by the image processing ECU 5 matches the actual object, the parking assist ECU 21 has the actual inter-vehicle distance. Based on a certain inter-vehicle distance S, the inter-vehicle distance from an object such as an obstacle can be appropriately determined, and normal parking assistance control can be executed. On the other hand, for example, even when the corresponding image selected by the image processing ECU 5 is different from the actual target object and the inter-vehicle distance S is different from the actual inter-vehicle distance, the parking assist ECU 21 intentionally deviates from the actual inter-vehicle distance. Based on the small inter-vehicle distance S, the inter-vehicle distance from an obstacle or the like is determined, and parking support control is executed. In this case, since the parking assistance ECU 21 recognizes the inter-vehicle distance S that is smaller than the actual inter-vehicle distance, the parking assistance control is performed so that the distance to the obstacle on the route becomes larger than usual. Therefore, a sufficient distance from an obstacle on the route is secured and parking assistance is performed, so that the safety of the vehicle is improved. That is, the vehicle control can be optimally supported.

  As mentioned above, although the best mode for carrying out the present invention has been described using the embodiment, the present invention is not limited to such an embodiment and is within the scope not departing from the gist of the present invention. Various modifications and substitutions can be made to the above-described embodiment.

  For example, in the above embodiment, the image processing ECU 5 selects one corresponding image from a plurality of corresponding image candidates based on the control mode of the host vehicle such as PCS control, ACC control, and parking support control. May be. That is, you may combine 1st Embodiment, 2nd Embodiment, and 3rd Embodiment which were mentioned above.

  Specifically, when the image processing ECU 5 detects a plurality of corresponding image candidates, the image processing ECU 5 transmits the inter-vehicle distance S of the corresponding image farther from the host vehicle to the collision corresponding ECU among the corresponding image candidates. The inter-vehicle distance S of the corresponding image closer to the own vehicle may be transmitted to the inter-vehicle ECU 19 and the parking assist ECU 21. In this case, the inter-vehicle distance S of the corresponding image transmitted to the inter-vehicle ECU 19 may be smaller than the inter-vehicle distance S of the corresponding image transmitted to the parking assist ECU 21. Thereby, a corresponding image is selected according to the control mode of the host vehicle such as PCS control, ACC control, and parking support control, and an appropriate inter-vehicle distance S is calculated.

  In the above embodiment, the cameras 3R and 3L are provided in the left-right direction, and the image processing ECU 5 calculates the inter-vehicle distance S between the object and the host vehicle based on the parallax Δd in the left-right direction. The camera may be provided in the vertical direction, and the image processing ECU 5 may calculate the inter-vehicle distance S between the target object and the host vehicle based on the vertical parallax.

  In the above embodiment, the image processing ECU 5 corresponds to the distance calculation means and the corresponding point selection means described in the claims.

  The present invention can be used for, for example, a vehicle image processing apparatus used for PCS control, ACC control, parking support control, and the like. The appearance, weight, size, running performance, etc. of the vehicle to be mounted are not limited.

It is a block diagram which shows an example of the system configuration | structure of the vehicle image processing apparatus which concerns on the 1st thru | or 3rd Embodiment of this invention. It is a figure which shows an example of the state arrange | positioned so that a camera may face the vehicle front in the right side and left side of a vehicle front part. It is a figure which shows an example of the parallax with the reference image image | photographed with the right camera, and the search image image | photographed with the left camera. It is a figure which shows an example of the distance (distance between vehicles) between a target object and the own vehicle. It is a flowchart which shows an example of the processing flow of the vehicle image processing apparatus which concerns on 1st Embodiment. It is a flowchart which shows an example of the process flow of the image processing apparatus for vehicles which concerns on 2nd Embodiment. It is a flowchart which shows an example of the processing flow of the image processing apparatus for vehicles which concerns on 3rd Embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Image processing apparatus for vehicles 3R, 3L Camera 5 Image processing ECU
7 Collision ECU
19 Vehicle ECU
21 Parking Assist ECU

Claims (2)

A vehicle image processing device mounted on a vehicle that performs pre-crash control and inter-vehicle distance control,
Stereo image acquisition means for acquiring a stereo image around the host vehicle;
Corresponding point detecting means for detecting corresponding points between the stereo images acquired by the stereo image acquiring means;
A distance calculation unit that calculates a parallax based on the corresponding point detected by the corresponding point detection unit, and calculates a distance between the host vehicle and an object around the host vehicle based on the parallax;
With
When a plurality of corresponding point candidates are detected by the corresponding point detection unit, the control unit that performs the pre-crash control is compared with the distance that is output to the control unit that performs the inter-vehicle distance control. It is characterized by outputting a long distance,
An image processing apparatus for a vehicle. A vehicle image processing apparatus mounted on a vehicle that performs parking support control and inter-vehicle control,
Stereo image acquisition means for acquiring a stereo image around the host vehicle;
Corresponding point detecting means for detecting corresponding points between the stereo images acquired by the stereo image acquiring means;
A distance calculation unit that calculates a parallax based on the corresponding point detected by the corresponding point detection unit, and calculates a distance between the host vehicle and an object around the host vehicle based on the parallax;
With
When a plurality of corresponding point candidates are detected by the corresponding point detection unit, the control unit that performs the parking assist control is compared with the distance that is output to the control unit that performs the inter-vehicle control. It is characterized by outputting a long distance,
An image processing apparatus for a vehicle.

Images