JP6169949B2 - Image processing device - Google Patents

Description

  The present invention relates to a technical field relating to an image processing apparatus that calculates a parallax and a distance of a subject based on a pair of captured images obtained by stereo imaging of the traveling direction of a host vehicle.

JP 2009-8539 A

  For example, in a vehicle system that recognizes the environment outside the vehicle, there is a vehicle system that recognizes a subject based on a captured image obtained by performing stereo imaging with a pair of cameras in the traveling direction of the host vehicle. In addition, in such a system for recognizing a subject existing in the traveling direction of the host vehicle, there is a system that generates three-dimensional position information (position information in real space) of the subject including information on the distance to the subject. . Specifically, captured images with different viewpoints are obtained by the above-described stereo imaging, the parallax of the subject is calculated based on the captured images, and the three-dimensional position information of the subject is generated based on the parallax.

The above three-dimensional position information is a point in space when the point directly below the center of the pair of cameras is the origin, the X axis is the direction connecting the pair of cameras, the Y axis is the vertical direction, and the Z axis is the front-back direction. It is represented by (X, Y, Z). The three-dimensional position information is generated based on the triangulation method. Specifically, the pixel on the image is represented by coordinates (i, j), the parallax is dp, the distance between the pair of cameras is CD, 1 When the viewing angle per pixel is PW, the mounting height of the pair of cameras is CH, and the i coordinate and the j coordinate on the image at the infinity point in front of the camera are IV and JV, respectively, It is obtained by coordinate transformation represented by [Formula 3].
X = CD / 2 + Z × PW × (i-IV) [Formula 1]
Y = CH + Z × PW × (j−JV) [Formula 2]
Z = CD / {PW × (dp−DP)} [Formula 3]
However, “DP” in the above [Expression 3] is a parallax offset value, which is determined when the pair of cameras are attached to the vehicle, and is fixed to the determined value thereafter.

Since the parallax offset value DP is a value determined based on a state in which the camera is attached to the host vehicle in a predetermined manner, for example, the camera attachment state is predetermined due to vibration or heat generated by use of the vehicle. If it changes from the state, the value of the distance Z cannot be calculated correctly by the above [Equation 3].
Therefore, as a technique for correcting an error in the parallax offset value DP caused by such a change in the camera mounting state over time, the present applicant previously proposed a correction technique described in Patent Document 1 above. ing. Specifically, in Patent Document 1, a total of four values (size b ₁ , b ₂ , parallax dp ₁ , dp ₂ ) of the apparent size of the preceding vehicle and the parallax dp are acquired at two different times, respectively. A parallax error ε that is an error between the parallax dp actually calculated from the value and the true parallax dp ^* is calculated, and the parallax offset value DP is corrected by the parallax error ε.

  According to the technique described in Patent Document 1, it is not necessary to image a predetermined test subject in correcting the parallax offset value DP, and appropriate correction can be performed under the actual use environment of the vehicle.

However, in Patent Document 1 described above, since the process related to the correction of the parallax offset value DP is executed for the entire period in which the preceding vehicle is detected, a relatively long time is required for the correction of the parallax offset value DP. It may take. Specifically, in Patent Document 1, after obtaining the apparent size b ₁ and the parallax dp ₁ of the preceding vehicle at a certain time, the apparent size after waiting for the parallax dp of the preceding vehicle to change to some extent. B ₂ and the parallax dp ₂ are acquired. At this time, for example, if the host vehicle is cruising with almost no speed difference from the preceding vehicle, the value of the parallax dp does not change, so the magnitude b ₂ and the parallax dp ₂ Cannot be obtained, and it may take a long time to correct the parallax offset value DP.

  In addition, since the entire period of the period in which the preceding vehicle is detected is targeted, the process related to the calculation of the correction value is continued even while the distance change from the preceding vehicle cannot be expected, resulting in a processing burden. There is also a risk of increase.

  SUMMARY OF THE INVENTION Accordingly, an object of the present invention is to overcome the above-described problems and to shorten the time required for correction and reduce the processing load related to the correction in a system for correcting a calculation formula for calculating a distance from parallax. .

  First, the image processing apparatus according to the present invention calculates a distance to the subject based on the parallax of the subject calculated based on a pair of captured images obtained by stereo imaging of the traveling direction of the host vehicle. Means, a preceding vehicle detecting means for detecting a preceding vehicle existing in the traveling direction of the host vehicle based on the captured image and the distance calculated by the distance calculating means, and a preceding vehicle detected by the preceding vehicle detecting means. Correction means for correcting a calculation formula for calculating the distance from the parallax based on the parallax corresponding to the vehicle and the change in the apparent size of the preceding vehicle, the correction means comprising: Based on the result of determining whether or not the brake lamp is lit, a process for calculating a correction value for correcting the calculation formula is started.

  As a result, it is possible to start the process related to the calculation of the correction value in response to the case where the preceding vehicle is in a decelerating state, that is, when the change in the parallax and the apparent magnitude of the preceding vehicle can be reliably expected.

Second, in the above-described image processing apparatus according to the present invention, the correction unit converts a frame image determined that the real space upper width between the left and right brake lamps of the preceding vehicle is equal to or smaller than a predetermined value to the correction value. It is desirable to exclude from the processing target related to the calculation.
As a result, for example, there is a high possibility that the object detected as the left and right brake lamps that are lit is not the left and right brake lamps of the same preceding vehicle, or there is a possibility that some detection error factor has occurred. Frame images are excluded from processing targets related to calculation of correction values.

Thirdly, in the above-described image processing apparatus according to the present invention, the correction means calculates a correction value based on a frame image that has been determined that the difference in width between the left and right brake lamps of the preceding vehicle is greater than or equal to a predetermined value. It is desirable to exclude them from such processing.
As a result, for example, a frame image with low reliability in which the width of the left and right brake lamps is different in a case where only one brake lamp is lit, etc., is excluded from the processing related to the calculation of the correction value. Is done.

Fourth, in the above-described image processing apparatus according to the present invention, the correction means calculates a correction value for a frame image that has been determined that the difference between the vertical positions of the left and right brake lamps of the preceding vehicle is greater than or equal to a predetermined value. It is desirable to exclude it from the target of processing related to.
As a result, if the host vehicle or the preceding vehicle is turning and has a roll angle, or if a red light emitter other than the brake lamp is erroneously detected on one of the left and right brake lamps, or there may be some detection error factor For example, a frame image with low reliability such as the above is excluded from processing targets related to calculation of correction values.

Fifth, in the above-described image processing apparatus according to the present invention, the correction unit excludes a frame image when it is determined that the host vehicle is turning from a target of processing related to the calculation of the correction value. Is desirable.
As a result, the frame image obtained when the host vehicle is not directly facing the rear surface of the preceding vehicle and the apparent size of the preceding vehicle is inaccurate is excluded from the processing related to the calculation of the correction value. .

  According to the present invention, in a system that corrects a calculation formula for calculating a distance from parallax, it is possible to shorten the time required for correction and reduce the processing load related to the correction.

It is the figure which showed the structure of the vehicle control system of embodiment. It is a figure for demonstrating the image processing performed in embodiment. It is the figure which illustrated the relationship between the width between brake lamps and brake lamp parallax. It is a flowchart of the main process in a parallax offset value correction process. It is a flowchart of a brake lamp parallax and a brake lamp width accumulation process. It is explanatory drawing about the information calculated by brake lamp parallax and the brake lamp width | variety accumulation process. It is a flowchart of an approximate line calculation process.

<1. Overall system configuration>
FIG. 1 shows a configuration of a vehicle control system 1 including an image processing apparatus as an embodiment according to the present invention. In FIG. 1, only the configuration of the main part according to the present invention is extracted and shown from the configuration of the vehicle control system 1.
The vehicle control system 1 includes an imaging unit 2, an image processing unit 3, a memory 4, a driving support control unit 5, a display control unit 6, an engine control unit 7, a transmission control unit 8, and a brake control unit provided for the host vehicle. 9, sensor / operator 10, display unit 11, engine-related actuator 12, transmission-related actuator 13, brake-related actuator 14, and bus 15.

  The image processing unit 3 executes predetermined image processing related to recognition of the environment outside the vehicle based on captured image data obtained by the imaging unit 2 imaging the traveling direction of the host vehicle (front in this example). The image processing by the image processing unit 3 is performed using a memory 4 which is a non-volatile memory, for example. Details of the internal configuration of the imaging unit 2 and specific processing executed by the image processing unit 3 will be described later.

The driving support control unit 5 is configured by a microcomputer including, for example, a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), and the like. Various control processes for driving support (hereinafter referred to as “driving support control process”) are executed based on detection information, operation input information, and the like obtained by the operators 10.
The driving support control unit 5 is connected to each control unit of the display control unit 6, the engine control unit 7, the transmission control unit 8, and the brake control unit 9 which are also configured by a microcomputer via a bus 15. Data communication can be performed with the control unit. The driving support control unit 5 instructs a necessary control unit among the above control units to execute an operation related to driving support.

The sensors / operators 10 comprehensively represent various sensors and operators provided in the host vehicle. The sensor / operator 10 includes a speed sensor 10A that detects the speed of the host vehicle, a brake switch 10B that is turned on / off in response to the operation / non-operation of the brake pedal, and an accelerator opening based on the depression amount of the accelerator pedal. There is an accelerator opening sensor 10C that detects the degree, a steering angle sensor 10D that detects the steering angle, a yaw rate sensor 10E that detects the yaw rate, and a G sensor 10F that detects the acceleration. Although not shown, other sensors include, for example, an engine speed sensor, an intake air amount sensor that detects the intake air amount, and an intake air amount that is interposed in the intake passage and supplied to each cylinder of the engine. A throttle opening sensor that detects the opening of the throttle valve, a water temperature sensor that detects the cooling water temperature indicating the engine temperature, an outside air temperature sensor that detects the temperature outside the vehicle, and the like are also provided.
In addition, as an operator, an ignition switch for instructing start / stop of the engine, selection of an automatic transmission mode / manual transmission mode in an AT (automatic transmission) vehicle, and an instruction for up / down in the manual transmission mode are given. There are a select lever for performing, a display changeover switch for switching display information in an MFD (Multi Function Display) provided in the display unit 11 described later, and the like.

  The display unit 11 comprehensively represents various meters such as speedometers and tachometers and MFDs provided in a meter panel installed in front of the driver, and other display devices for presenting information to the driver. . Various information such as the total travel distance of the host vehicle, the outside air temperature, and instantaneous fuel consumption can be displayed on the MFD simultaneously or by switching.

  The display control unit 6 controls the display operation by the display unit 11 based on a detection signal from a predetermined sensor in the sensor / operator 10, operation input information by the operator, and the like. For example, based on an instruction from the driving support control unit 5, a predetermined alert message can be displayed on the display unit 11 (for example, a predetermined area of the MFD) as part of driving support.

The engine control unit 7 controls various actuators provided as the engine-related actuator 12 based on a detection signal from a predetermined sensor in the sensor / operator 10 or operation input information by the operator. As the engine-related actuator 12, for example, various actuators related to engine driving such as a throttle actuator that drives a throttle valve and an injector that performs fuel injection are provided.
For example, the engine control unit 7 performs engine start / stop control according to the operation of the ignition switch described above. The engine control unit 7 also controls the fuel injection timing, the fuel injection pulse width, the throttle opening, and the like based on detection signals from predetermined sensors such as an engine speed sensor and an accelerator opening sensor 10C.

The transmission control unit 8 controls various actuators provided as the transmission-related actuator 13 based on a detection signal from a predetermined sensor in the sensor / operator 10 or operation input information by the operator. As the transmission-related actuator 13, there are provided various transmission-related actuators such as a control valve that performs shift control of an automatic transmission and a lockup actuator that locks up a lockup clutch.
For example, when the automatic shift mode is selected by the select lever described above, the transmission control unit 8 performs a shift control by outputting a shift signal to the control valve according to a predetermined shift pattern.
Further, when the manual shift mode is set, the transmission control unit 8 performs a shift control by outputting a shift signal according to a shift up / down instruction from the select lever to the control valve.

  The brake control unit 9 controls various actuators provided as the brake-related actuators 14 based on detection signals from predetermined sensors in the sensors / operators 10 and operation input information by the operators. As the brake-related actuator 14, for example, various brake-related actuators such as a hydraulic pressure control actuator for controlling the hydraulic pressure output from the brake booster to the master cylinder and the hydraulic pressure in the brake fluid piping are provided. For example, when an instruction to turn on the brake is issued from the driving support control unit 5, the brake control unit 9 controls the hydraulic pressure control actuator to brake the host vehicle. The brake controller 9 calculates the slip ratio of the wheel from the detection information of a predetermined sensor (for example, the axle rotation speed sensor or the vehicle speed sensor 10A), and applies the hydraulic pressure by the hydraulic control actuator according to the slip ratio. By reducing the pressure, so-called ABS (Antilock Brake System) control is realized.

<2. Image processing executed in this embodiment>
The image processing executed in this embodiment will be described with reference to FIG.
In FIG. 2, in order to describe image processing, the internal configuration of the imaging unit 2 and the memory 4 shown in FIG. 1 are shown together with the configuration of the image processing unit 3. First, the imaging unit 2 for obtaining captured image data used for image processing will be briefly described.
The imaging unit 2 includes a first camera unit 20-1, a second camera unit 20-2, an A / D converter 21-1, an A / D converter 21-2, and an image correction unit 22. .
Each of the first camera unit 20-1 and the second camera unit 20-2 includes a camera optical system and an image sensor such as a CCD (Charge Coupled Device) or a CMOS (Complementary Metal Oxide Semiconductor). A subject image is formed on the imaging surface of the imaging element by the system, and an electrical signal corresponding to the amount of received light is obtained in pixel units by the imaging element.
The first camera unit 20-1 and the second camera unit 20-2 are installed so as to enable distance measurement by a so-called stereo imaging method. That is, they are arranged so that a plurality of captured images with different viewpoints can be obtained. The first camera unit 20-1 and the second camera unit 20-2 in this example are arranged at a predetermined interval in the vehicle width direction in the vicinity of the upper part of the windshield of the host vehicle. The optical axes of the first camera unit 20-1 and the second camera unit 20-2 are parallel, and the focal lengths are the same. Also, the frame periods are synchronized and the frame rates are the same.

  The electrical signal obtained by the imaging device of the first camera unit 20-1 is sent to the A / D converter 21-1, and the electrical signal obtained by the imaging device of the second camera unit 20-2 is sent by the A / D converter 21. -2 and A / D conversion is performed respectively. As a result, a digital image signal (image data) representing a luminance value with a predetermined gradation in pixel units is obtained. In the present embodiment, the first camera unit 20-1 and the second camera unit 20-2 can obtain color captured images of R (red), G (green), and B (blue). As the luminance value, values corresponding to R, G, and B are obtained. In the case of this example, the number of effective pixels of the image sensor included in the first camera unit 20-1 and the second camera unit 20-2 is 1280 pixels in the horizontal direction × 960 pixels in the vertical direction.

  The image correction unit 22 includes image data (hereinafter referred to as “first captured image data”) based on an image captured by the first camera unit 20-1 obtained via the A / D converter 21-1, A / D Image data (hereinafter referred to as “second captured image data”) based on an image captured by the second camera unit 20-2 obtained via the D converter 21-2 is input. For example, the image correction unit 22 corrects a shift caused by an error in the attachment positions of the first camera unit 20-1 and the second camera unit 20-2 for each of the first captured image data and the second captured image data. Use affine transformation or the like. The image correction unit 22 also corrects the luminance value including noise removal for each of the first captured image data and the second captured image data.

  The first captured image data and the second captured image data obtained by the imaging unit 2 are recorded and held in the memory 4 by the image processing unit 3.

Here, in the case of this example, the imaging unit 2 generates two types of images, a bright image and a dark image, having different shutter speeds for each of the first captured image data and the second captured image data. The dark image is an image captured at a shutter speed faster than that of the bright image, and is an overall dark image compared to the bright image. The first camera unit 20-1 and the second camera unit 20-2 perform imaging with the shutter speed changed every frame period (for example, 1/60 seconds) and alternately output a bright image and a dark image. Thereby, as the 1st picked-up image data obtained by the 1st camera part 20-1, and the 2nd picked-up image data obtained by the 2nd camera part 20-2, a bright image and a dark image are respectively for every frame period. It will be switched alternately.
The dark image is used when a parallax offset value correction processing unit 3E described later detects lighting of a brake lamp of a preceding vehicle or the like. A bright image is used for generation of three-dimensional position information, which will be described later, and object detection processing by the preceding vehicle detection processing unit 3D.

The image processing unit 3 is configured by, for example, a microcomputer, and executes various types of image processing based on the first captured image data and the second captured image data according to the activated program.
In FIG. 2, various types of image processing executed by the image processing unit 3 are shown in blocks for each function. As shown in the figure, the image processing unit 3 is roughly divided into functions, and the three-dimensional position information generation processing unit 3A, the lane detection processing unit 3B, the lane model formation processing unit 3C, the preceding vehicle detection processing unit 3D, and the parallax offset value A correction processing unit 3E is provided.

The three-dimensional position information generation processing executed by the three-dimensional position information generation processing unit 3A is a process of generating three-dimensional position information based on the first captured image data and the second captured image data as bright images held in the memory 4 It becomes. Specifically, in the three-dimensional position information generation process, corresponding points between the first captured image data and the second captured image data (that is, a pair of image data captured in stereo) are detected by pattern matching, and the detected correspondence is detected. This is a process of calculating a coordinate shift between points as a parallax dp and generating information on the position of the corresponding point in the real space as three-dimensional position information by using the parallax dp.
In calculating the coordinate shift as the parallax dp as described above, one of the first captured image data and the second captured image data is determined in advance as a “reference image” and the other as a “comparison image”. The comparison image is generated as an image having a larger number of pixels in the horizontal direction than the reference image in order to allow the calculation of the parallax dp for the object located at the horizontal end on the reference image.

Here, the three-dimensional position information is based on a point directly below the center of the pair of cameras (the first camera unit 20-1 and the second camera unit 20-2) as the origin, and the X-axis and the vertical direction in the direction connecting the pair of cameras. Is the information expressed as a point (X, Y, Z) in space when the Y axis is taken as the Z axis in the front-rear direction.
The respective values of X, Y, and Z as the three-dimensional position information are expressed as pixel coordinates (i, j) when the axis parallel to the horizontal direction in the reference image is i-axis and the axis parallel to the vertical direction is j-axis. ), The distance between the pair of cameras is CD, the viewing angle per pixel is PW, the mounting height of the pair of cameras is CH, and the i-coordinate and j-coordinate on the reference image at the infinity point in front of the camera are When IV and JV are set, they are obtained by coordinate transformation represented by the following [Formula 1] to [Formula 3].
X = CD / 2 + Z × PW × (i-IV) [Formula 1]
Y = CH + Z × PW × (j−JV) [Formula 2]
Z = CD / {PW × (dp−DP)} [Formula 3]
“DP” in [Expression 3] is also referred to as a vanishing point parallax, an infinite distance corresponding point, or the like, but in short, the parallax dp between corresponding points between the reference image and the comparison image, and in the real space up to the corresponding point Is a value determined so as to satisfy the above [Equation 3]. Hereinafter, this “DP” is expressed as “parallax offset value DP”.

  The lane detection processing executed by the lane detection processing unit 3B includes a reference image as a bright image (that is, one of the first captured image data and the second captured image data set in advance) and the above three-dimensional position. Based on the three-dimensional position information generated by the information generation process (including the distance Z for each pixel as a corresponding point), this is a process for detecting a lane formed on the road surface on which the host vehicle travels. Specifically, in the lane detection process, first, a lane candidate point is detected on the reference image based on the luminance value of each pixel of the reference image and the distance Z of each pixel in the real space, and based on the detected lane candidate point. The left and right lane positions of the host vehicle are detected. For example, a search is performed by offsetting a horizontal line having a width of one pixel on the reference image by one pixel in the left-right direction, and the luminance differential value (= edge strength) of each pixel is a threshold value based on the luminance value of each pixel of the reference image. Pixels that satisfy the conditions that change greatly as described above are detected as lane candidate points. This process is sequentially performed while offsetting the horizontal line to be searched for by one pixel width upward, for example, from the lower side of the reference image. Thereby, a lane candidate point is detected in each of the right region and the left region of the host vehicle.

  The lane model formation processing executed by the lane model formation processing unit 3C is based on the information on the left and right lane candidate points detected by the above lane detection and is in a three-dimensional space defined by the X, Y, and Z axes. This is a process for forming a lane model. Specifically, the three-dimensional position (X, Y, Z) of the lane candidate point detected by the lane detection unit is linearly approximated by, for example, the least square method to form a lane model in the three-dimensional space. With the lane model formed in this way, the height information of the road surface on which the host vehicle travels can also be obtained.

The preceding vehicle detection process executed by the preceding vehicle detection processing unit 3D is a process for detecting a preceding vehicle existing ahead of the host vehicle based on the reference image as a bright image and the three-dimensional position information. In the preceding vehicle detection process, first, an object detection process is performed based on the three-dimensional position information, and an object present in the image is detected including information on the distance Z to the object. For example, in this object detection process, a distance image is generated in which each corresponding point detected in the previous parallax dp calculation process is associated with the value of the distance Z on the image, and the distance image is partitioned in the vertical direction. Divide into multiple vertical areas, create a distance histogram that represents the distance distribution in the vertical direction of the image (j direction) for each vertical area, and the distance Z (corresponding point) where the frequency is maximum exists in that vertical area This is the representative distance of the target object. Then, for each corresponding point having the maximum frequency for which the representative distance is obtained, pixel ranges that are regarded as the same object are grouped from the relationship such as the distance Z and direction to each adjacent corresponding point, and exist in the image. Specify the range of each object. Thereby, an object existing in the image is detected including information on the distance Z to the object.
Here, the distance image is obtained sequentially for each frame. In the preceding vehicle detection process, by monitoring the information of the distance Z of the detected object over a plurality of frames, an object that exists on the traveling path of the own vehicle and moves at a predetermined speed in substantially the same direction as the own vehicle. Extract as a preceding vehicle. At this time, in order to suppress erroneous detection of an object other than the vehicle, pattern matching using the reference image (for example, pattern matching based on a feature point of the vehicle such as a brake lamp portion) is also performed.
When the preceding vehicle is detected, the preceding vehicle detection information includes the preceding vehicle distance (= inter-vehicle distance), the preceding vehicle speed (= change ratio of the inter-vehicle distance + own vehicle speed), and the preceding vehicle acceleration (= differential of the preceding vehicle speed). Value).
Note that the method of the preceding vehicle detection process is the same as the method disclosed in Japanese Patent Application Laid-Open No. 2012-66759. For details, refer to this document.

  The preceding vehicle detection processing unit 3D can detect the replacement of the preceding vehicle. Replacement of a preceding vehicle means that a preceding vehicle that was being detected has left the front of the host vehicle and a vehicle in front of the preceding vehicle is newly detected as a preceding vehicle, When another vehicle is detected as a preceding vehicle instead of the preceding vehicle being detected, such as when another vehicle interrupts the host vehicle and the other vehicle is newly detected as a preceding vehicle. Means.

  In the parallax offset value correction processing executed by the parallax offset value correction processing unit 3E, the distance Z is calculated from the parallax dp based on the parallax of the preceding vehicle detected by the preceding vehicle detection processing unit 3D and the change in the apparent size. This is processing for correcting the calculation formula (that is, the previous [Formula 3]). As described above, since the parallax offset value DP in [Equation 3] has an error from the optimum value corresponding to the actual camera mounting state due to the change with time of the camera mounting state. Is to correct.

Here, prior to description of specific processing contents executed as the parallax offset value correction processing, an outline of the parallax offset value DP correction processing performed in this example will be described.
The point of performing correction based on the change in the parallax and the apparent size of the preceding vehicle detected in the captured image is the same as in the case of Patent Document 1 mentioned above. Regarding the size, information on the width between the left and right brake lamps of the preceding vehicle is used.
Here, the “apparent size of the preceding vehicle” means the size of the preceding vehicle that can be seen with reference to the vehicle position.

Further, the correction processing of this example is a total of four values of the apparent magnitude and the parallax of the preceding vehicle (magnitudes b ₁ , b ₂ , parallax dp ₁₎ at two different times as in the correction method described in Patent Document _1. , Dp ₂ ) at least, and the concept of calculating the correction value of the parallax offset value DP from these values is common, but as in Patent Document 1, the parallax error ε is obtained and the parallax offset value is calculated based on the parallax error ε. It does not correct DP.
In this example, at least two different time points are acquired, and a total of four values of the brake lamp width and the parallax of the preceding vehicle (brake lamp parallax described later in this example) as the apparent size of the preceding vehicle are acquired, Approximate straight lines based on coordinates of at least two points obtained when the four values are developed on a two-dimensional coordinate space in which the horizontal axis / vertical axis are the apparent size (width between brake lamps) / parallax, respectively (y = ax + b) is obtained, and the parallax offset value DP is corrected by the value of the intercept (b) of the approximate straight line.

  FIG. 3 is a diagram illustrating the relationship between the brake lamp width and the brake lamp parallax (the parallax dp of the brake lamp of the preceding vehicle). In the figure, the ◆ marks indicate the combinations of [brake lamp width, brake rake lamp parallax] obtained at a plurality of different time points, with the brake lamp width / brake lamp parallax on the vertical and horizontal axes, respectively. It is expressed in a dimensional coordinate space.

  The width between brake lamps of the preceding vehicle increases as the preceding vehicle becomes closer to the own vehicle, and conversely decreases as the distance from the own vehicle increases. Similarly, the value of the brake lamp parallax increases as the preceding vehicle becomes closer to the host vehicle, and decreases as the distance from the host vehicle increases. For this reason, it can be considered that a proportional relationship is generally established between the brake lamp width and the brake lamp parallax, and the relationship between the brake lamp width and the brake lamp parallax is expressed by a linear function as “y = ax + b”. be able to.

Here, when the preceding vehicle is at an infinite point, the value of the brake lamp width is “0”. Further, when the preceding vehicle is at the infinity point, the value of the brake lamp parallax is ideally “0”. For this reason, ideally, the value of the intercept b should be “0” in the linear function representing the relationship between the brake lamp width and the brake lamp parallax as described above.
However, when the camera mounting state changes with time as described above, the value of the brake lamp parallax when the preceding vehicle is at the infinity point does not become “0”, and the required offset to the value of the intercept b Occurs.

  Therefore, in the correction process of this example, the brake lamp width and the brake lamp parallax are acquired at least at two different times as described above, and the brake lamp width and the brake light are calculated based on the acquired brake lamp width and the brake lamp parallax. A straight line representing the relationship with the lamp parallax is obtained by primary approximation, and the parallax offset value DP is corrected based on the value of the intercept b of the obtained approximate straight line. Specifically, the parallax offset value DP is corrected by “DP + (− b)” using a value obtained by inverting the sign of the intercept b as a correction value.

<3. Correction Processing of Embodiment>
Based on the outline of the correction process described above, a specific process procedure executed by the image processing unit 3 as the parallax offset value correction process will be described below.
FIG. 4 is a flowchart of the main process in the parallax offset value correction process executed by the image processing unit 3. Note that the main process shown in FIG. 4 is repeatedly executed each time a set of bright and dark images is obtained.

First, in step S101, the image processing unit 3 determines whether or not the host vehicle is turning. Whether the host vehicle is turning is determined based on detection signals from the steering angle sensor 10D and the yaw rate sensor 10E.
If a positive result is obtained that the host vehicle is turning, the image processing unit 3 ends the main process shown in FIG.

  Here, when the host vehicle is turning, since the host vehicle does not face the rear surface of the preceding vehicle, the apparent size of the preceding vehicle becomes inaccurate. Therefore, the frame image obtained in such a case It is not desirable to calculate a correction value for correcting the parallax offset value DP based on the above. Therefore, the accuracy of the correction value is improved by terminating the main processing as described above and excluding the frame image from the processing target.

On the other hand, if a negative result is obtained that the host vehicle is not turning, the image processing unit 3 proceeds to step S102 and determines whether there is a preceding vehicle and there is no replacement. That is, it is determined whether or not a preceding vehicle has been detected by the above-described preceding vehicle detection process and a change in the preceding vehicle has not been detected.
If a negative result is obtained that the condition that there is a preceding vehicle and there is no replacement is not satisfied, the image processing unit 3 proceeds to step S107, clears the accumulated data, and ends the main processing shown in FIG. That is, the accumulated data of the brake lamp parallax and the brake lamp width accumulated in step S103 to be described next is cleared (erased), and the main process ends.
As a result, it is possible to prevent an incorrect correction value from being calculated based on the brake lamp parallax and brake lamp width data accumulated for different preceding vehicles, thereby improving the correction accuracy.

  Further, when a positive result is obtained that the condition that there is a preceding vehicle and there is no replacement is satisfied, the image processing unit 3 proceeds to step S103 and performs brake lamp parallax and brake lamp width accumulation processing. Execute. The accumulation process in step S103 is a process for accumulating data on the brake lamp parallax and the brake lamp width for the preceding vehicle at a plurality of times, which will be described in detail later.

  In response to the execution of the accumulation process in step S103, the image processing unit 3 determines whether or not the data accumulation number has reached a predetermined number or more in step S104. That is, it is determined whether or not the number of accumulated brake lamp parallax and brake lamp width data has reached a predetermined number or more. As understood from the above description, in calculating the approximate straight line representing the relationship between the brake lamp width and the brake lamp parallax, the number of samples of the brake lamp width and the brake lamp parallax may be at least two, The “predetermined number” may be set to at least “2” or more. In order to improve the correction accuracy, it is desirable that the “predetermined number” is larger.

  If a negative result is obtained that the data accumulation number is not equal to or greater than the predetermined number, the image processing unit 3 finishes the main process shown in FIG.

  On the other hand, when an affirmative result is obtained that the number of accumulated data is greater than or equal to the predetermined number, the image processing unit 3 proceeds to step S105 and executes approximate straight line calculation processing. The approximate straight line calculation process of step S105 is a process of calculating an approximate straight line representing the relationship between the brake lamp width and the brake lamp parallax based on the data accumulated in the process of step S103. Details of the approximate straight line calculation process will be described later.

  In response to the calculation of the approximate line in step S105, the image processing unit 3 executes a process of correcting the parallax offset value based on the value of the intercept b of the approximate line in step S106, and the main process shown in FIG. 4 is completed. . As described above, in the correction processing of this example, a value obtained by inverting the sign of the intercept b is calculated as a correction value, and the parallax offset value DP is corrected by “DP + (− b)”.

FIG. 5 is a flowchart of the brake lamp parallax and brake lamp width data accumulation process (S103).
First, in step S201, the image processing unit 3 determines whether the left and right brake lamps of the preceding vehicle are lit. The discrimination process in step S201 is performed using the above-described dark image. Specifically, it is determined whether or not there are two or more image portions where the red luminance value is equal to or greater than a predetermined value within the same image range as the preceding vehicle in the dark image.

If a positive result is obtained that the left and right brake lamps of the preceding vehicle are lit, the image processing unit 3 advances the process to step S202. That is, processing for accumulating data on the brake lamp parallax and the brake lamp width, which will be described later, is executed.
On the other hand, if a negative result is obtained that the left and right brake lamps are not lit, the image processing unit 3 ends the process shown in FIG. That is, the process for accumulating data of the brake lamp parallax and the brake lamp width described below is not executed.

Here, when accurately obtaining the approximate straight line as described above, the brake lamp parallax corresponding to the parallax of the preceding vehicle and the brake lamp width corresponding to the apparent size of the preceding vehicle are obtained as discrete values as possible. It is desirable to be able to do it. That is, it is desirable that the distance can be acquired when the inter-vehicle distance from the preceding vehicle changes relatively large.
In view of this point, in the present embodiment, in calculating the correction value of the parallax offset value DP, it is determined whether the brake lamp of the preceding vehicle is lit as described above, and the correction value is calculated based on the determination result. The processing related to the calculation is started (or not started). As a result, it becomes possible to start the process related to the calculation of the correction value in response to the case where the preceding vehicle is in a deceleration state, that is, when the parallax and the apparent magnitude of the preceding vehicle can be reliably expected. It is possible to efficiently acquire discrete values for the parallax and the apparent size.

If an affirmative result is obtained in step S201, the image processing unit 3 calculates left and right brake lamp widths, left and right brake lamp center coordinates, an upper image width between brake lamps, and a brake lamp parallax in step S202.
Here, the left and right brake lamp widths are the left brake lamp width WL and the right brake lamp width WR of the preceding vehicle as shown in FIG. For these widths WL and WR, the widths of the image portions (red light emission portions) in which the red luminance value detected in the same range as the image range of the preceding vehicle in the dark image is equal to or greater than a predetermined value are obtained.
Depending on the type of the preceding vehicle, a red light emitter may be detected between the left and right brake lamps such as a high-mount stop lamp. When three or more red light emitting portions are detected in the image range of the preceding vehicle, only the red light emitting portions at the left and right ends of them are extracted and defined as left and right brake lamps.

Further, the left and right brake lamp center coordinates are represented by black circles in FIG. 6, and are the center coordinates of the red light emitting portions as the left and right brake lamps obtained as described above.
Further, the upper image width between brake lamps is represented by “BW” in FIG. 6, and means the width between the left and right brake lamp center coordinates obtained as described above.
The brake lamp parallax is the parallax dp of the brake lamp of the preceding vehicle as described above. In this example, the average value of the parallax dp of each of the left and right brake lamps is calculated as the brake lamp parallax. At this time, the parallax dp of each of the left and right brake lamps is obtained by obtaining the center coordinates of the left brake lamp and the center coordinates of the right brake lamp on both the first captured image data side and the second captured image data side. It calculates | requires as dp, respectively.
The brake lamp parallax may be the parallax dp of one of the left and right brake lamps.

In step S202, the image processing unit 3 calculates the left and right brake lamp width, the left and right brake lamp center coordinates, the brake lamp image upper width, and the brake lamp parallax in step S203. In step S204, it is determined whether or not the actual width between the brake lamps is equal to or smaller than a predetermined value. The real space width between brake lamps is obtained by converting the image width between brake lamps calculated in step S202 into a width in real space (three-dimensional space). Then, in the determination process in step S204, it is determined whether or not the actual width between the brake lamps is, for example, 1 m or less.
If an affirmative result is obtained in step S204 that the upper space between the brake lamps is equal to or smaller than the predetermined value, the image processing unit 3 ends the process shown in FIG. That is, for a frame image in which it is determined that the actual space between the brake lamps is equal to or less than a predetermined value, calculation of a brake lamp parallax and brake lamp width for calculating a correction value, which will be described later (S207), and accumulation (S208). Is not performed, and is excluded from the processing related to the calculation of the correction value.

  Here, if the width between the objects detected as the left and right brake lamps that are lit is too narrow, there is a high possibility that the left and right objects are not the left and right brake lamps of the same preceding vehicle. Alternatively, although the left and right brake lamps of the same preceding vehicle are detected, there may be some temporary detection error factors due to, for example, an external environment or a malfunction in the system. Therefore, as described above, the correction of the correction value based on the low-reliability frame image is performed by excluding the frame image that is determined to have the upper real space width between the left and right brake lamps equal to or less than the predetermined value. This prevents the breakage and improves the accuracy of the correction value.

If a negative result is obtained in step S204 that the upper space between the brake lamps is not less than or equal to the predetermined value, the image processing unit 3 proceeds to step S205 and determines the difference between the left and right brake lamp widths (WL in FIG. 6). It is determined whether or not (difference between WR and WR) is a predetermined value or more, for example, 2 pixels or more.
If a positive result is obtained that the difference between the left and right brake lamp widths is greater than or equal to a predetermined value, the image processing unit 3 finishes the process shown in FIG. That is, the current frame image is excluded from the processing target related to the correction value calculation.

Usually, since the brake lamp is provided in a symmetrical shape, the detection size should be equal if the host vehicle is facing the rear surface of the preceding vehicle. On the other hand, if there is a difference between the left and right brake lamp widths, it can be considered that some abnormality has occurred. For example, in a case where only one brake lamp of the preceding vehicle is relatively strongly lit in the early morning or evening time zone, the detection size of one brake lamp may be large. Alternatively, the width of the left and right brake lamps may be different not only in the external environment such as the above-mentioned direct light but also in the case where some temporary detection error factor is generated due to a problem on the system. The difference in the width of the left and right brake lamps means that at least one of the left and right brake lamp center coordinates described above is inaccurate, and accordingly, the value of the width between the brake lamps corresponding to the apparent size of the preceding vehicle is also obtained. Become inaccurate.
Therefore, the correction value is calculated based on the unreliable frame image by excluding the frame image determined as having the difference between the left and right brake lamp widths of the preceding vehicle being equal to or greater than the predetermined value as described above. This prevents the breakage and improves the accuracy of the correction value.

If a negative result is obtained in step S205 that the difference between the left and right brake lamp widths is not equal to or greater than the predetermined value, the image processing unit 3 proceeds to step S206, and the difference between the vertical positions of the left and right brake lamps is equal to or greater than the predetermined value. It is determined whether or not. Here, the vertical positions of the left and right brake lamps are the above-described vertical positions (j coordinates) of the center coordinates as indicated by “JL” and “JR” in FIG. In step S206, it is determined whether or not the difference between the vertical positions is, for example, 4 pixels or more.
When a positive result is obtained that the difference between the left and right brake lamp widths is equal to or greater than a predetermined value, the image processing unit 3 ends the process illustrated in FIG. That is, the current frame image is excluded from the processing target related to the correction value calculation.

  Here, when there is a difference in the detected vertical position of the left and right brake lamps, the host vehicle or the preceding vehicle is turning, has a roll angle, and the brake lamp is arranged obliquely in the captured image, or the same preceding vehicle There is a possibility that other than the brake lamp of the preceding vehicle is erroneously detected, such as one of the pair of left and right objects that should have been detected as the left and right brake lamps is a red light emitter other than the brake lamp. Alternatively, although the left and right brake lamps of the same preceding vehicle are detected, there may be some temporary detection error factors due to an external environment or a problem on the system. Since it is not desirable to calculate the correction value in these cases, by excluding the frame image determined that the difference in the vertical position is equal to or greater than the predetermined value as described above from the processing target related to the calculation of the correction value, The calculation of the correction value based on the frame image with low reliability is prevented, and the accuracy of the correction value is improved.

If a negative result is obtained in step S206 that the difference between the vertical positions of the left and right brake lamps is not equal to or greater than a predetermined value, the image processing unit 3 proceeds to step S207 and calculates the correction value calculation brake lamp parallax and the correction value calculation. Calculate the brake lamp width. The brake lamp parallax for calculating the correction value is calculated as follows using the value of the brake lamp parallax obtained in step S202.

Brake lamp parallax for calculating correction value = base line length / (pixel pitch * focal length) / brake lamp parallax + parallax offset value DP [Equation 4]

Also, the brake lamp width for calculating the correction value is calculated as follows using the value of the brake lamp image width obtained in step S202.

Correction value calculation brake lamp width = (base line length * brake lamp image width) / {base line length / (pixel pitch * focal length)} / distance [Equation 5]

However, in [Expression 5], “distance” means the distance to the preceding vehicle. For example, the value of the distance Z calculated in [Expression 3] using the brake lamp parallax value obtained in step S202 is used. Use. Or you may use the value of the preceding vehicle distance calculated | required by the preceding vehicle detection process mentioned above.

  Here, the correction value calculation brake lamp parallax and the correction value calculation brake lamp parallax calculated as described above are both values incorporating the parallax offset value DP (between the correction value calculation brake lamps). Since the width is finally divided by the distance, the parallax offset value DP is taken in). In this embodiment, as described above, the values of the brake lamp parallax and the brake lamp width (the image width between brake lamps) are not used as they are as the values of the brake lamp parallax and the brake lamp width. These values are converted into a form in which the parallax offset value DP is taken in, and an approximate straight line is calculated (that is, a correction value is calculated) in step S105 using the converted values. Thereby, the accuracy of the correction value can be improved as compared with the case of calculating an approximate straight line using the values of the brake lamp parallax and the brake lamp width in units of the number of pixels as they are.

  For confirmation, as will be understood from the fact that the distance between the brake lamps for calculating the correction value is finally divided by the distance in [Equation 5], the preceding vehicle moves away from the own vehicle. The value decreases as the value approaches, and the value increases as the value approaches. Therefore, the width between the brake lamps for calculating the correction value still represents the apparent size of the preceding vehicle.

  The image processing unit 3 calculates the correction value calculation brake lamp parallax and the correction value calculation brake lamp parallax in step S207, and calculates the correction value calculation brake lamp parallax and the correction value calculated in step S208. The process of accumulating the calculation brake lamp width in the memory 4 is executed, and the process shown in FIG.

FIG. 7 is a flowchart of the approximate line calculation process (S105).
The image processing unit 3 resets the processing count value n to 0 in step S301 shown in FIG. 7 in response to obtaining an affirmative result that the data accumulation number is equal to or greater than the predetermined number in the previous step S104. In step S302, an approximate straight line is calculated by the least square method based on the accumulated data. That is, based on the data of [correction value calculating brake lamp parallax, correction value calculating brake lamp parallax] accumulated by the processing of the previous step S208, an approximate straight line ( y = ax + b) is calculated by the method of least squares.

  In response to the calculation of the approximate straight line in step S302, the image processing unit 3 proceeds to step S303, and among the accumulated data, data having an excessive residual with respect to the approximate straight line and data in which the value of the brake lamp width deviates. Execute the process to remove. That is, among the accumulated data of [correction value calculation brake lamp parallax, correction value calculation brake lamp width], data in which the residual with respect to the approximate straight line calculated in step S302 is excessive (residual is predetermined). Data greater than or equal to the value) and data for which the value of the correction value calculating brake lamp width deviates from the value of the correction value calculating brake lamp width of the other data. At this time, whether or not the value of the correction value calculation brake lamp width is deviated is obtained, for example, by calculating an average value of a plurality of accumulated correction value calculation brake lamp width values and calculating a difference from the average value. Is determined by whether or not is equal to or greater than a predetermined value.

  In subsequent step S304, the image processing unit 3 recalculates the approximate line by the least square method based on the accumulated data. That is, the approximate straight line is recalculated based on the accumulated data excluding the data removed in step S303.

In the next step S305, the image processing unit 3 determines whether or not the processing count value n is greater than or equal to the upper limit value N. If a negative result is obtained, the processing count value is incremented in step S306 (n ← n + 1), and the process returns to step S303. On the other hand, when a positive result is obtained in step S305, the image processing unit 3 ends the process illustrated in FIG.
As a result, the data removal process in step S303 and the approximate straight line recalculation process in step S304 are repeated until the processing count value n reaches the upper limit value N. By repeating this, a straight line with higher accuracy can be calculated as the approximate straight line, and thus the accuracy of the correction value can be improved. In this example, the upper limit value N = 1, and the approximate straight line is calculated three times by the least square method.

In response to the execution of the approximate straight line calculation process of step S105 as described above, the image processing unit 3 executes the correction process of step S106 described above. In the process of step S106 in this case, the parallax offset value DP is corrected based on the value of the intercept b of the approximate straight line finally calculated by the above-described iterative process.
Thereby, the error of the parallax offset value DP caused by the change with time of the camera mounting state can be corrected appropriately. That is, the calculation formula for calculating the distance Z from the parallax dp can be appropriately corrected.

<4. Summary of Embodiment>
As described above, the image processing apparatus (image processing unit 3) according to the present embodiment is able to reach the subject based on the parallax of the subject calculated based on a pair of captured images obtained by stereo imaging of the traveling direction of the host vehicle. A preceding vehicle existing in the traveling direction of the host vehicle is detected based on the distance calculation means (three-dimensional position information generation processing unit 3A) for calculating the distance of the vehicle, and the captured image and the distance calculated by the distance calculation means Based on the parallax corresponding to the preceding vehicle detected by the preceding vehicle detecting means and the apparent magnitude of the preceding vehicle, based on the change in the apparent size of the preceding vehicle. And a correction unit (parallax offset value correction processing unit 3E) that corrects the calculation formula for calculating the distance.
And the said correction | amendment means starts the process which concerns on calculation of the correction value for correct | amending the said calculation formula based on the result of having determined whether the brake lamp of the said preceding vehicle is lighting.

Accordingly, it is possible to start the process related to the calculation of the correction value in response to the case where the preceding vehicle is in a deceleration state, that is, when the parallax and the apparent magnitude of the preceding vehicle can be reliably expected.
Therefore, in the system that corrects the calculation formula for calculating the distance from the parallax, the time required for the correction can be shortened and the processing load related to the correction can be reduced.

Further, in the image processing apparatus according to the present embodiment, the correction unit relates to the calculation of the correction value for a frame image determined that the real space width between the left and right brake lamps of the preceding vehicle is equal to or less than a predetermined value. Exclude from processing.
As a result, for example, there is a high possibility that the object detected as the left and right brake lamps that are lit is not the left and right brake lamps of the same preceding vehicle, or there is a possibility that some detection error factor has occurred. Frame images are excluded from processing targets related to calculation of correction values.
Accordingly, the accuracy of the correction value can be improved, and the calculation accuracy of the distance Z can be improved.

Furthermore, in the image processing apparatus according to the present embodiment, the correction unit performs processing related to calculation of the correction value for a frame image that is determined that the difference in width between the left and right brake lamps of the preceding vehicle is greater than or equal to a predetermined value. Exclude from the target.
As a result, for example, a frame image with low reliability in which the width of the left and right brake lamps is different in a case where only one brake lamp is lit, etc., is excluded from the processing related to the calculation of the correction value. Is done.
Accordingly, the accuracy of the correction value can be improved, and the calculation accuracy of the distance Z can be improved.

Furthermore, in the image processing apparatus according to the present embodiment, the correction means relates to the calculation of the correction value for a frame image that has been determined that the difference between the vertical positions of the left and right brake lamps of the preceding vehicle is greater than or equal to a predetermined value. Exclude from processing.
As a result, if the host vehicle or the preceding vehicle is turning and has a roll angle, or if a red light emitter other than the brake lamp is erroneously detected on one of the left and right brake lamps, or there may be some detection error factor For example, a frame image with low reliability such as the above is excluded from processing targets related to calculation of correction values.
Accordingly, the accuracy of the correction value can be improved, and the calculation accuracy of the distance Z can be improved.

In addition, in the image processing apparatus according to the present embodiment, the correction unit excludes the frame image when it is determined that the host vehicle is turning from the processing target related to the calculation of the correction value.
As a result, the frame image obtained when the host vehicle is not directly facing the rear surface of the preceding vehicle and the apparent size of the preceding vehicle is inaccurate is excluded from the processing related to the calculation of the correction value. .
Accordingly, the accuracy of the correction value can be improved, and the calculation accuracy of the distance Z can be improved.

<5. Modification>
While the embodiments of the present invention have been described above, the present invention is not limited to the specific examples described above, and various modifications can be considered.
For example, the conditions of the frame image excluded from the processing target related to the calculation of the correction value are not limited to those exemplified above. For example, when the front / rear pitch amount of the host vehicle is greater than or equal to a predetermined value, there is a concern that the calculation accuracy of the correction value may be reduced. Therefore, the frame image obtained in such a case can be excluded from the processing target. . The front / rear pitch amount of the host vehicle can be obtained based on the detection signal of the G sensor 10F. In addition, since the forward pitch is generated when the host vehicle is in the brake ON state, the frame image when it is determined that the brake switch 10B is ON may be excluded from the processing target.
Further, instead of determining whether the host vehicle is turning (step S101), it may be determined whether there is a curve within a predetermined distance on the traveling path of the host vehicle.

  Moreover, the correction method of the distance calculation formula illustrated above is only an example, and a correction method for calculating the parallax error ε as a correction value as described in Patent Document 1, for example, can be employed. In addition, without calculating the brake lamp parallax and the brake lamp width for calculating the correction value, the approximate straight line is calculated using the values of the brake lamp parallax and the brake lamp width (image width between brake lamps) in units of pixels. You can also

  DESCRIPTION OF SYMBOLS 1 ... Vehicle control system, 2 ... Imaging part, 3 ... Image processing part, 3A ... Three-dimensional position information generation processing part, 3D ... Prior vehicle detection processing part, 3E ... Parallax offset value correction processing part

Claims (5)

Distance calculating means for calculating a distance to the subject based on the parallax of the subject calculated based on a pair of captured images obtained by stereo imaging of the traveling direction of the host vehicle;
Preceding vehicle detection means for detecting a preceding vehicle existing in the traveling direction of the host vehicle based on the captured image and the distance calculated by the distance calculation means;
Correction means for correcting a calculation formula for calculating the distance from the parallax based on the parallax corresponding to the preceding vehicle detected by the preceding vehicle detection means and a change in the apparent size of the preceding vehicle. ,
The correction means includes
An image processing apparatus that starts processing related to calculation of a correction value for correcting the calculation formula based on a result of determining whether or not a brake lamp of the preceding vehicle is lit. The correction means includes
The image processing apparatus according to claim 1, wherein a frame image that is determined to have an upper real space width between the left and right brake lamps of the preceding vehicle that is equal to or less than a predetermined value is excluded from processing targets related to the calculation of the correction value. The correction means includes
The image processing apparatus according to claim 1, wherein a frame image that is determined that a difference in width between left and right brake lamps of the preceding vehicle is equal to or greater than a predetermined value is excluded from processing targets related to calculation of the correction value. The correction means includes
The frame image determined that the difference between the vertical positions of the left and right brake lamps of the preceding vehicle is greater than or equal to a predetermined value is excluded from the processing related to the calculation of the correction value. Image processing device. The correction means includes
The image processing apparatus according to any one of claims 1 to 4, wherein a frame image when it is determined that the host vehicle is turning is excluded from processing targets related to the calculation of the correction value.

Images