JP2021180441A - Image processing device - Google Patents

Abstract

To provide an image processing device that can detect an abnormality in an in-vehicle camera even when the change in the edge amount of image data of the camera is small at the time of performing failure diagnosis of the camera in real time while driving.SOLUTION: An image processing device includes an image detection unit that detects and identifies a target to be detected from image input data of an in-vehicle camera which is a detection target, an image restoration unit that restores an image from the image input data, an overcorrection detection unit that detects an overcorrected portion on the basis of the restoration result, and an abnormality determination unit that makes an abnormality determination from the detection result of the target and the detection result of the overcorrected portion, and detects an abnormality in the camera while driving by detecting an overcorrected area.SELECTED DRAWING: Figure 1

Description

The present application relates to an image processing device which is mounted on a vehicle and performs a failure diagnosis of a camera in real time while driving.

When an image processing device detects the target of another vehicle, pedestrian, bicycle, two-wheeled vehicle or other obstacle on the road from the image taken by the camera mounted on the vehicle in recent years, the target exists at a predetermined distance. When it is judged, a warning is issued, and an accident is prevented by automatically operating the brake or the handle. Here, the "target" is a term used in the field of radar system or traffic control system, and is a term corresponding to a target. In the image processing device mounted on the vehicle, if the camera breaks down, there is a high possibility that the control becomes unstable. Therefore, an extremely reliable device is required.

Therefore, it has been proposed to detect an abnormality in the camera by detecting an edge of an image captured by the camera as an imaging means and comparing the amount of the detected edge with a reference value. (See, for example, Patent Document 1).

Japanese Unexamined Patent Publication No. 9-11321

"Restoration of Space Variant Degraded Images Using Wavelet Transform" https://search.ieice.org/bin/pdf_link.php?category=D2&lang=J&year=1995&fname=j78-d2_12_1821&abst=

However, in the conventional image processing apparatus according to Patent Document 1, since the abnormality of the camera is detected by the amount of edges, the change in the amount of edges due to the abnormality of the camera is compared with the change in the amount of edges in the captured image. Since it is small, it is difficult to detect abnormalities in the camera, and there is a problem that detection of abnormalities in focus blur of the camera, which has a serious effect on image processing, may be delayed, and there is a problem that measures must be taken for this. rice field.

The present application has been made to solve the above-mentioned problems, and when a failure diagnosis of an in-vehicle camera is performed in real time while driving, an abnormality of the camera is detected even if the change in the edge amount of the image data of the camera is small. It is an object of the present invention to provide an image processing apparatus capable of performing.

The image processing apparatus disclosed in the present application includes an image detection unit that detects a target to be detected from image data of a camera and outputs a detection result, an image restoration unit that outputs a restored image restored from the image data, and an image restoration unit. It is characterized by including an overcorrection detection unit that detects an overcorrection region from the restored image, and an abnormality determination unit that determines the presence or absence of an abnormality in the camera by the overcorrection region.

According to the image processing apparatus disclosed in the present application, by determining the abnormality of the camera by the area of the overcorrection of the restored image, the abnormality of the camera can be detected even if the edge amount of the image data of the camera changes. There is an effect.

It is a functional block diagram which shows the whole structure of the image processing system including the image processing apparatus which concerns on Embodiment 1. FIG. It is a schematic block diagram of the image processing apparatus which concerns on Embodiment 1. FIG. It is a flow figure which shows the abnormality determination procedure by the image processing apparatus which concerns on Embodiment 1. FIG. It is a figure which shows the region of the overcorrection in Embodiment 1. FIG. It is a figure which shows the area of the target object to be detected in Embodiment 1. FIG. It is a figure which shows the area of the target which does not detect anomaly in Embodiment 1. FIG. It is a figure which shows the product inspection process in Embodiment 1. FIG.

Embodiment 1.
FIG. 1 is a functional block diagram showing an overall configuration of an image processing system including the image processing apparatus according to the first embodiment. FIG. 2 is a schematic configuration diagram of the image processing apparatus according to the first embodiment. FIG. 3 is a flow chart showing an abnormality determination procedure by the image processing apparatus according to the first embodiment.

Next, with reference to FIG. 1, the entire configuration of the image processing system including the image processing apparatus according to the first embodiment will be described. The image processing device 1 has an image detection unit 11 and an image detection unit 11 that detect and identify a pedestrian, another vehicle, and a target target including a white line from the image input data of the camera 10 to be detected. The detection result output unit 12 that outputs the detection result of the target, the first image restoration unit 13a and the second image restoration unit 13b that restore the image from the image input data of the camera, and the first image restoration unit. An object by the first overcorrection detection unit 14a, the second overcorrection detection unit 14b, and the image detection unit 11 that detect the overcorrected portion based on the restoration results by the 13a and the second image restoration unit 13b, respectively. The abnormality determination unit 15 that determines an abnormality from the detection result of the mark and the detection result of the overcorrection portion by the first overcorrection detection unit 14a and the second overcorrection detection unit 14b, and the determination result of the abnormality determination unit 15 are output. It is composed of a determination result output unit 16.

As shown in FIG. 2, each functional unit 11 to 16 included in the image processing device 1 is realized by a processing device 80, a storage device 81, an input device 82, an output device 83, and a display device 84.

Here, the processing unit 80 is a CPU (also referred to as a Central Processing Unit, a central processing unit, a microprocessor, a microcomputer, a processor, or a DSP) that executes a program stored in the storage device 81 even if it is dedicated hardware. It may be.

When the processing device 80 is dedicated hardware, the processing device 80 corresponds to, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC, an FPGA, or a combination thereof. .. Image detection unit 11, detection result output unit 12, first image restoration unit 13a, second image restoration unit 13b, first overcorrection detection unit 14a, second overcorrection detection unit 14b, and determination result output unit 16. The functions of each part may be realized by the processing device 80, or the functions of each part may be collectively realized by the processing device 80.

When the processing device 80 is a CPU, the image detection unit 11, the detection result output unit 12, the first image restoration unit 13a, the second image restoration unit 13b, the first overcorrection detection unit 14a, and the second overcorrection detection. The functions of each part of the unit 14b and the determination result output unit 16 are realized by software, firmware, or a combination of software and firmware. The software and firmware are described as a processing program and stored in the storage device 81. The processing device 80 realizes the functions of each unit by reading and executing the processing program stored in the storage device 81. That is, from the processing step for capturing camera image data from the camera 10, the processing step for detecting image detection from the acquired camera image data, and the camera image data when the image processing device 1 is executed by the processing device 80. A processing program that results in execution of a processing process for image restoration, a processing process for detecting excessive correction due to image restoration, a processing process for determining the presence or absence of an abnormality, and a processing process for outputting data processing results to an external device. A storage device 81 for storing the data is provided. Further, these processing programs include an image detection unit 11, a detection result output unit 12, a first image restoration unit 13a, a second image restoration unit 13b, a first overcorrection detection unit 14a, and a second overcorrection detection. It can also be said that the procedure or method of the unit 14b and the determination result output unit 16 is executed by the computer. Here, the storage device 81 corresponds to, for example, a non-volatile or volatile semiconductor memory such as RAM, ROM, flash memory, EPROM, EEPROM, a magnetic disk, a flexible disk, an optical disk, a compact disk, a mini disk, a DVD, or the like. do.

Some of the functions of the image detection unit 11, the first image restoration unit 13a, the second image restoration unit 13b, the first overcorrection detection unit 14a, and the second overcorrection detection unit 14b are dedicated. It may be realized by hardware, and partly by software or firmware. For example, the functions of the first image restoration unit 13a and the second image restoration unit 13b are realized by the processing device 80 as dedicated hardware, and the first overcorrection detection unit 14a and the second overcorrection detection are realized. The function of the unit 14b can be realized by the processing device 80 reading and executing the program stored in the storage device 81.

As described above, the processing device 80 can realize each of the above-mentioned functions by hardware, software, firmware, or a combination thereof.

In addition to storing the program that executes the above-mentioned processing step, the storage device 81 uses the camera image data acquired from the camera 10 and the data processed by the abnormality determination unit 15 to compare with the reference value to determine the abnormality. Stores the error judgment result of the camera.

Further, the input device 82 periodically acquires the image data of the camera output from the camera 10 at a predetermined time by the image detection unit 11, the first image restoration unit 13a, and the second image restoration unit 13b. .. The output device 83 corresponds to the detection result output unit 12 and the determination result output unit 16, and outputs the processing result to the external device. The display device 84 appropriately displays the status and the like executed by the processing device 80.

In FIG. 1, the camera 10 is mounted on a vehicle and outputs camera image data of images taken from the front, side, or rear of the vehicle toward the image processing device 1.
The image detection unit 11 detects a pedestrian, another vehicle, and a target to be detected by the white line from the image data of the camera output by the camera 10, and the area (position, size) on the image of the detected target. Information including) is output to the detection result output unit 12 as a detection result. The detection result output unit 12 outputs the target detection result to the external device. Further, the image detection unit 11 provides the abnormality determination unit 15 with the information of the region on the image of the target that does not detect the abnormality of the “sky” as the detection result. The abnormality determination unit 15 detects the presence or absence of an abnormality in the camera based on the image detection unit 11, the first overcorrection detection unit 14a, the second overcorrection detection unit 14b, and the vehicle state information from the external device. The abnormality detection result of the camera indicating the position and size of the object to be detected that has been detected as an abnormality on the screen is output to the external device through the determination result output unit 16.

Next, with reference to FIG. 1, the image restoration unit that restores the image of the target to be detected from the image data captured by the camera mounted on the vehicle, and the restored image data are overcorrected. The processing contents of each part of the overcorrection detection unit that detects the damaged area and the abnormality determination unit that determines the presence or absence of an abnormality in the specified area on the screen from the position and time of the detected target will be described.

First, in the first image restoration unit 13a and the second image restoration unit 13b, the deterioration parameter σ is the image of the image data of the deteriorated camera in _{the first deterioration parameter σ 1} and the second deterioration parameter σ _{2, respectively.} To restore. Regarding image restoration, the method of Non-Patent Document 1 is referred to.

ここで、ｘ，ｙは画素の座標を、σ^２は分散を、σ（∝ｒ）は半値幅をそれぞれ示す。 As the deterioration parameter σ, the PSF (Point Spread Function) h (x, y), which is the impulse response of the (degraded) system to the point light source, is half of the model when the Gaussian function of Eq. (1) is used. The price range σ will be used.

Here, x and y indicate the coordinates of the pixel, σ ² indicates the variance, and σ (∝r) indicates the half width.

The first image restoration unit 13a and the second image restoration unit 13b restore the image data f (x, y) before being deteriorated by h (x, y) by the image data g (x, y) of the camera. do. g (x, y) is represented by the equation (2).

Here, n (x, y) is a noise component, but will be omitted below for the sake of simplicity.
The first image restoration unit 13a and the second image restoration unit 13b perform image restoration using the wavelet transform. The image restoration unit first converts the camera image data g (x, y) and PSF h (x, y) into G (a, b), H (a, b) (b = (x, y)), respectively. Wavelet transform. G (a, b), H (a, b) and the wavelet transform F (a, b) of the image data f (x, y) before deterioration are in the relationship of the equation (3).

F (a, b) is obtained from the equation (3) (formula (4)), and the image data f (x, y) before deterioration is restored by the wavelet inverse transformation.

式（７）のσ＞σ_Ｆのとき、Ｆ（ａ，ｂ）は、式（６）のＦ（ａ，ｂ）＝１のウェーブレット逆変換 ｆ（ｘ，ｙ）＝（点光源の画像）を越える補正が行われ、過大な補正が行われる（図４を参照。）。 The first overcorrection detection unit 14a and the second overcorrection detection unit 14b make excessive corrections in the restored image data output from the first image restoration unit 13a and the second image restoration unit 13b, respectively. Detects the damaged area (position of pixels on the screen) and outputs it.
For example, the wavelet transform of image data of the PSF of the degradation parameter _{σ F H F (a, b} ) and, when the _{G (a, b) = H} F (a, b) in equation (4), the deterioration of this When the image is restored as image data deteriorated by the PSF of the parameter σ, F (a, b) becomes the following equations (5) to (7) _{depending on the magnitude relationship between σ and σ F.} Here, H _{| σ-σF |} (a, b) is the wavelet transform of the PSF of the _{deterioration parameter | σ-σ F |.}

_{When σ> σ F in} the equation (7), F (a, b) is a wavelet inverse transform of F (a, b) = 1 in the equation (6) f (x, y) = (image of a point light source). A correction exceeding the above is performed, and an excessive correction is performed (see FIG. 4).

When the camera image data has an abnormality of defocusing of the deterioration parameter σ or more, even if the image is restored by the deterioration parameter σ, the excessive correction of the equation (7) is not performed.
It can be seen that no abnormality has occurred in the area on the screen where the excessive correction has been performed (see FIG. 6).
The first overcorrection detection unit 14a and the second overcorrection detection unit 14b are on the image in which the region on the image on which the overcorrection is performed is restored by reverse wavelet transforming F (a, b). It is detected by the negative pixel value generated by ringing that does not originally exist in the image.

The abnormality determination unit 15 is a region on the screen where the excessive correction detected by the first overcorrection detection unit 14a and the second overcorrection detection unit 14b has been performed, that is, a region where an abnormality in defocusing has not occurred. From the time when it is detected and the position and time of the target detected by the image detection unit 11, it is determined whether or not there is an abnormality in the designated area on the screen.

The designated area on the screen in the present embodiment is defined as the area Pi (i = 0 to 8) shown in FIG. The abnormality determination unit 15 counts the count values c [n] [i] (n = 1,2, where n =) for each region Pi, for each of the first overcorrection detection unit 14a and the second overcorrection detection unit 14b. 1 represents the first overcorrection detection unit 14a, and n = 2 represents the second overcorrection detection unit 14b.) A counter is provided, and the count value c is provided every fixed time (every time the camera image data is input). [n] [i] are subtracted (counted down) by the subtraction value Δc [n] [i], respectively. Further, the abnormality determination unit 15 counts the count value c [when the area Pi includes the area where the excessive correction, which is the output of the first overcorrection detection unit 14a and the second overcorrection detection unit 14b, is performed. Set n] [i] to the initial values s [n] [i] (n = 1,2), respectively. Also, when the system starts up, the count value c [n] [i] is set to s [n] [i].
Further, the abnormality determination unit 15 outputs the determination result of abnormality to the determination result output unit 16 when the count value c [1] [i] or c [2] [i] becomes '0' or less. ..

式（８）において、｜Pi｜は領域Piの面積(画素数)、｜Si_target｜は検知された検知対象の物標の領域のうちの領域Piに含まれる領域の面積(画素数)である（図５を参照。）。｜Si_mask｜は、検知された異常検出しない物標の領域のうちの領域Piに含まれる領域の面積(画素数)である（図６を参照。）。d_target[n][i]は、検知対象の物標が検知されたときの最大の減算値である。d_mask[n][i]は、異常検出しない物標が検知されないときの最大の減算値である。d₀[n][i]は、最小の減算値である。減算値Δc[n][i]は、d₀[n][i]を最小として、検知された検知対象の物標と異常検出しない物標の領域が領域Piに含まれる割合で決める。 Further, the abnormality determination unit 15 uses the subtraction value Δc [n] [i] as the equation (8).

In the equation (8), | Pi | is the area of the region Pi (number of pixels), and | Si _target | is the area of the region included in the region Pi of the detected target target regions (number of pixels). Yes (see Figure 5). | Si _mask | is the area (number of pixels) of the area (number of pixels) included in the area Pi in the area of the target that is detected and the abnormality is not detected (see FIG. 6). d _target [n] [i] is the maximum subtraction value when the target to be detected is detected. d _mask [n] [i] is the maximum subtraction value when a target that does not detect anomalies is not detected. d ₀ [n] [i] is the smallest subtraction value. The subtraction value Δc [n] [i] is determined by the ratio that the area of the detected target target and the area of the target not detected abnormally are included in the area Pi with _{d 0 [n] [i] as the minimum.}

The initial values of the counter s [n] [i], d _target [n] [i], d _mask [n] [i], d ₀ [n] [i] are the position on the screen and the deterioration parameter σ _n. Select according to (n = 1,2). For example, since the central area of the screen detects a small target to be detected in the distance, there is a concern that deterioration of detection performance due to camera abnormality may cause a serious situation, and it is necessary to detect deterioration at an early stage. , Decrease the initial value s [1] [i] of the counter of the deterioration parameter σ ₁ (<σ _{2 ), d target} [1] [i], d _mask [1] [i], d ₀ [1] [ i] should be increased. On the other hand, the area at the top of the screen does not detect the target to be detected, so d _target [1] [i] = 0, d _mask [1] [i] = 0, d ₀ [1] [ It is possible to detect an abnormality with the deterioration parameter σ _{2 by setting i] = 0.}

The determination result output unit 16 outputs a determination result indicating the presence or absence of an abnormality in the abnormality determination unit 15.

Next, the processing method of abnormality detection in the first embodiment will be described with reference to FIGS. 3 to 6.
At system startup (t = 0) (step S101), s [n] [i] is set in the count values c [n] [i] of each region Pi (step S102).

The image detection unit 11 inputs the image data output from the camera 10 (step S103). The target to be detected is detected from the image data (step S104). Further, the image detection unit 11 detects a target that does not detect an abnormality from the camera image data (step S105).

The first image restoration unit 13a and the second image restoration unit 13b restore the camera image data with the _{designated deterioration parameters σ 1} and σ _{2, respectively (step S106).}

The first overcorrection detection unit 14a and the second overcorrection detection unit 14b perform excessive correction on a region having a negative pixel value that does not originally exist in the image, which is generated by ringing on the restored image. It is detected as a region (step S107).

The abnormality determination unit 15 is a region where the excessive correction detected by the first overcorrection detection unit 14a and the second overcorrection detection unit 14b has been performed, and a region of the target to be detected detected by the image detection unit. , It is determined whether or not there is an abnormality from the area where the abnormality is not detected (step S108).
When the subtraction value Δc [n] [i] in the equation (8) of the abnormality determination unit 15 is d _target [n] [i] = 0 and d _mask [n] [i] = 0, that is, Δc [n ] [I] = d ₀ [n] [i], the processing of the abnormality determination unit 15 will be described with reference to FIG.

If there is no abnormality in the camera 10, and the area Pi includes the overcorrected area output by the first overcorrection detection unit 14a or the second overcorrection detection unit 14b at time t1 in FIG. 4, the area Pi The initial value s [n] [i] is set in the count value c [n] [i] (step S109). Further, the process returns to step S103.
If there is an abnormality in the camera 10, there is no output in the overcorrection area, and the subtraction value Δc [n] [i] = d ₀ [n] [i] of the count value c [n] [i] continues ( Step S110, step S111), the count value c [n] [i] becomes '0' or less at time t2, and the abnormality of the camera 10 is detected (step S112). After that, the abnormality detection result is output (step S113).

検知対象の物標が検知されるまでは、｜Si_target｜=0で、Δc[n][i]=0であるので、c[n][i]は、一定値となる。
時刻t0において、領域Pi内に検知対象の物標が検知されると、カウント値c[n][i]が、式（９）の減算値Δc[n][i]ずつ減算される（ステップＳ１１０、ステップＳ１１１）。 When the subtraction value Δc [n] [i] of the equation (8) in the abnormality determination unit 15 is d ₀ [n] [i] = 0 and d _mask [n] [i] = 0, that is, the equation ( The processing of the abnormality determination unit 15 in the case of 9) will be described with reference to FIG.

_{Until the target} to be detected is detected, | Si target | = 0 and Δc [n] [i] = 0, so c [n] [i] is a constant value.
When the target to be detected is detected in the area Pi at time t0, the count value c [n] [i] is subtracted by the subtraction value Δc [n] [i] of the equation (9) (step). S110, step S111).

When there is no abnormality in the camera 10 and the area Pi includes the area where the excessive correction is performed, which is output by the first overcorrection detection unit 14a or the second overcorrection detection unit 14b at time t1, the area Pi is counted. The initial value s [n] [i] is set in the value c [n] [i] (step S109). After that, the process returns to the image data input of the camera (step S103).
If there is an abnormality in the camera 10, there is no output in the overcorrection area, the subtraction value Δc [n] [i] in Eq. (9) continues to subtract the count values c [n] [i], and the count is counted at time t2. When the values c [n] [i] become '0' or less, the abnormality of the camera 10 is detected (step S112). After that, the abnormality detection result is output (step S113).

Therefore, when there is no abnormality in the camera 10, the area of the target to be detected always includes the area of overcorrection and the count value is initialized, so that the count value does not become '0' or less. No anomaly detection is output. When there is an abnormality in the camera 10, since the area of the target to be detected does not include the area of overcorrection, the count value becomes ‘0’ or less at an early stage, and the abnormality can be detected in a short time.

領域Piに異常検出されない物標が含まれない場合、カウント値c[n][i]は、減算値Δc[n][i]=d_mask[n][i]で減算され、時刻t2において異常検出を出力する（ステップＳ１１０、ステップＳ１１１）。
時刻t0から、領域Piに異常検出しない物標が含まれる場合、領域Piの面積｜Pi｜と領域Piに含まれる異常検出しない物標の面積｜Si_mask｜の割合に応じて減算値Δc[n][i]が減少し、カウント値c[n][i]が‘0’となる時刻が先送りされる。そのため、領域Piが「空（そら）」の部分で過大補正が起きにくい場合に、異常と判定されることを防止する（ステップＳ１１２）。 When the subtraction value Δc [n] [i] of the equation (8) in the abnormality determination unit 15 is d _target [n] [i] = 0, d ₀ [n] [i] = 0, that is, the subtraction value. The processing of the abnormality determination unit 15 in the case where Δc [n] [i] is the equation (10) will be described with reference to FIG.

If the region Pi does not contain a target that is not detected abnormally, the count value c [n] [i] is subtracted by the subtraction value Δc [n] [i] = d _mask [n] [i], and at time t2. Output the abnormality detection (step S110, step S111).
From time t0, if it contains target object not abnormality detection in the region Pi, the area of the region Pi | Pi | and area target area which do not abnormality detection contained in the Pi | Si _mask | subtraction value in proportion to the .DELTA.c [ The time when n] [i] decreases and the count value c [n] [i] becomes '0' is postponed. Therefore, when the region Pi is “empty” and it is difficult for excessive correction to occur, it is prevented from being determined as abnormal (step S112).

The determination result output unit 16 outputs an abnormality detection result if there is an abnormality as a result of determination of the presence or absence of the abnormality determined by the abnormality determination unit 15 (step S113), and if it is normal, normal detection is performed. The result is output (step S114), and the process returns to the camera image data input (step S103).

In the above description, the case where there are a plurality of sets of the image restoration unit and the overcorrection detection unit has been described, but each may be a single unit.

As described above, in the first embodiment, since the abnormality of the camera is determined by the area of the overcorrection of the restored image, the abnormality of the camera can be detected even if the edge amount of the camera image data changes.

The region Pi may be one region. Further, the first image restoration unit 13a and the second image restoration unit 13b may use other image restoration methods such as a Wiener filter. Further, the first image restoration unit 13a and the second image restoration unit 13b may have different image restoration methods. Further, the first overcorrection detection unit 14a and the second overcorrection detection unit 14b may be different in accordance with them.

Further, in the first embodiment, the overcorrected region is detected by the negative pixel value generated in the restored image by ringing, but G (a, b) / H in the equation (4). (A, b) It may be detected by the characteristic of (function). For example, in the region b = (x, y) where the overcorrection is performed, the higher the frequency (a is smaller) of G (a, b) / H (a, b), the larger the value.

Further, in FIG. 1, the abnormality determination unit makes an abnormality determination only while the vehicle is running based on the vehicle state (speed) information, and when the state in which overcorrection is unlikely to occur while the vehicle is stopped continues. I try to prevent it from being judged as abnormal. Further, when the abnormality of the camera does not occur suddenly, the power consumption can be reduced by detecting the abnormality only for a certain period after the start-up.

Further, in the product inspection process shown in FIG. 7, as shown in FIG. 7A, for example, the inspection chart 20 of the pattern shown in FIG. 7B is photographed by the camera 10, and the abnormality determination unit is in each of the regions Pi. It is also possible to output the result of the abnormality determination of the product and use it to determine the quality of the product.

Also, although the present application describes exemplary embodiments, the various features, embodiments, and functions described in the embodiments are not limited to the application of a particular embodiment, but alone. , Or in various combinations can be applied to embodiments.
Therefore, innumerable variations not exemplified are envisioned within the scope of the techniques disclosed herein. For example, it is assumed that at least one component is modified, added or omitted.

Further, in the figure, the same reference numerals indicate the same or corresponding parts.

1 image processing device, 10 cameras, 11 image detection unit, 12 detection result output unit, 13a first image restoration unit, 13b second image restoration unit, 14a first overcorrection detection unit, 14b second overcorrection Detection unit, 15 Abnormality determination unit, 16 Judgment result output unit, 20 Inspection chart, 80 Processing device, 81 Storage device, 82 Input device, 83 Output device, 84 Display device

Claims (7)

An image detection unit that detects the target to be detected from the image data of the camera and outputs the detection result,
An image restoration unit that outputs the restored image restored from the image data, and
An overcorrection detection unit that detects an overcorrection area from the restored image,
An image processing device including an abnormality determination unit for determining the presence or absence of an abnormality in the camera by the overcorrection region. The first aspect of claim 1, wherein a plurality of sets of the image restoration unit and the overcorrection detection unit are provided, and the presence or absence of an abnormality in the camera is determined by the plurality of the overcorrection regions detected by the plurality of restored images. Image processing equipment. The image processing apparatus according to claim 1 or 2, wherein a designated deterioration parameter is used for the restoration of the restored image. Any one of claims 1 to 3, wherein the presence or absence of an abnormality in the camera is determined based on whether or not the overcorrection region is included in the plurality of regions set in the image data. The image processing apparatus according to. The image processing according to any one of claims 1 to 4, wherein the camera determines the presence or absence of an abnormality depending on whether or not the overcorrection region is detected within a specified fixed period. Device. The image processing apparatus according to claim 5, wherein the designated fixed period is changed depending on the area of the target area of the image data. The image processing apparatus according to claim 5, wherein the designated fixed period is changed depending on the area of an empty area of the image data.

Images