JPH07249128A - Vehicle image processing device - Google Patents

Abstract

(57) [Summary] [Purpose] Easy calculation, high-speed processing, and
(EN) An image processing device for a vehicle in which the hardware configuration scale can be reduced. [Structure] An edge detecting means shifts an image obtained by a CCD camera 11 by a predetermined number of pixels in horizontal and vertical directions,
Edge information is detected in the image state by comparing the brightness of the images before and after the shift.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a vehicular image processing apparatus which can be applied to, for example, a vehicular surroundings monitoring apparatus and which uses an image.

[0002]

2. Description of the Related Art Conventionally, a method of detecting a change amount (change rate) of an image signal using an edge extraction filter is generally used for detecting edge information by an image processing apparatus. FIG. 6A shows a Laplacian filter as a typical edge extraction filter (3 × 3 pixels) (Takeshi Yasui, “Computer image processing”, 18th).
See page 19). The edge is detected by obtaining the change amount of the vertical and horizontal luminance signals while moving the filter vertically and horizontally. The operation will be described below.
FIG. 6B shows an image region to be processed by the Laplacian filter, and a certain point (i,
j) The image signal of the center point (i, j) 61 is calculated by weighting the image signal of each point in the 3 × 3 square area centered on 61 with a coefficient as shown in FIG. Signal. That is, the image signal at the point (i, j) 61 is f
(I, j), the image signal after the arithmetic processing by the filter is g
Assuming that (i, j), g (i, j) is expressed by the following equation.

G (i, j) = (-1) * f (i-1, j-1) + (-1) * f (i, j-
1) + (-1) * f (i + 1, j-1) + (-1) * f (i-1, j) + (-8) * f (i, j)
+ (-1) * f (i + 1, j) + (-1) * f (i-1, j + 1) + (-1) * f (i, j + 1) +
(-1) * f (i + 1, j + 1)

FIG. 7A shows an example of an image taken by a CCD camera or the like. By performing the edge extraction processing by the above-mentioned Laplacian filter on this image, edge information as shown in FIG. 7B is detected.

Further, in the conventional vehicle periphery monitoring device for detecting a vehicle ahead of the own vehicle, in order to detect the vehicle, for example, the luminance signal is differentiated by utilizing the luminance difference between the vehicle image and the road surface, A vehicle image processing apparatus that extracts a vehicle candidate area by obtaining edge information is disclosed in Japanese Patent Publication No. 4-15134.
3 is disclosed.

The process for vehicle detection used in the vehicle image processing apparatus will be described below with reference to the drawings. FIG. 8 is a block diagram showing an example of processing. FIG. 9A is a front image including a preceding vehicle captured by a CCD camera or the like. First, edge information having a brightness difference is detected from the image 80 by the edge extraction processing 81, and the horizontal continuous edge extraction processing 82 is included in the detected edge information.
The edge information of the edge that is continuous in the horizontal direction is extracted by
By the vertical direction continuous edge extraction processing 83, edge information of edges continuous in the vertical direction on the image is extracted. FIG. 9B illustrates a state in which the edge information images extracted by the vertical continuous edge processing and the horizontal continuous edge processing are combined. In addition to this, the vehicle area extraction processing 84 further increases the vertical and horizontal 2
The vehicle 85 is extracted with the area where the type of continuous edge information is present as the vehicle area.

[0007]

However, in the edge detecting means used in the above-described conventional vehicle image processing apparatus, for example, when the size of the filter is 3 × 3 pixels, the image signal of one pixel is determined. However, there is a problem in that it is necessary to read out an image signal of up to 9 pixels for calculation, which takes time to process, and the scale of hardware also becomes large.

Further, in the conventional image processing apparatus for a vehicle, particularly in an actual driving scene, it is often difficult to detect the vehicle separately from the background because the background is complicated and changes greatly. For example, if there is a sign or a guardrail 92 as shown in FIG. 9A, in the conventional vehicle image processing apparatus, these backgrounds are likely to be erroneously extracted, and the accuracy of vehicle detection deteriorates. It was In addition, since it is necessary to process the image data of the entire screen of the image, it is difficult to increase the processing speed.

The present invention has been made for the purpose of solving the above-mentioned problems, and it is easy to perform the calculation, the processing speed can be increased, and the hardware image processing can be made small. The purpose is to provide a device.

[0010]

According to a first aspect of the present invention, there is provided an image processing apparatus for a vehicle, comprising: a photographing means for photographing the periphery of the vehicle; and an edge detecting means for detecting edge information from the image photographed by the photographing means. In the vehicular image processing device having the above, the edge detecting means shifts the image obtained by the photographing means in the horizontal / vertical direction by a predetermined number of pixels, and compares the brightness of the image before and after the shifting to obtain the image. The edge information is detected in the state.

A vehicle image processing apparatus according to a second aspect of the present invention is the vehicle image processing apparatus according to the first aspect, wherein the edge detecting means binarizes the detected edge information. , And a threshold value changing means for changing the threshold value when the binarization is performed by the binarizing means.

A vehicle image processing apparatus according to a third aspect of the present invention is the vehicle image processing apparatus according to the first or second aspect, wherein the edge information detected by the edge detecting means is vertically aligned on the image. It is provided with an edge projection means for integrating to obtain a one-dimensional graph.

A vehicle image processing apparatus according to a fourth aspect of the present invention is the vehicle image processing apparatus according to the first or second aspect, wherein in the edge information of the image state detected by the edge detecting means, It is provided with a laterally long edge projection means for detecting edge information representing a laterally long edge and integrating the laterally long edge information in the vertical direction on the image to obtain a one-dimensional graph.

A vehicle image processing apparatus according to a fifth aspect of the present invention is the vehicle image processing apparatus according to the third or fourth aspect, further comprising a change amount detecting means for detecting a change amount of the one-dimensional graph, When the change amount obtained by the change amount detecting means is equal to or more than a predetermined value, the vehicle recognizing means determines that the change amount is a vehicle.

A vehicle image processing apparatus according to a sixth aspect of the present invention is the vehicle image processing apparatus according to any one of the third to fifth aspects, wherein the one-dimensional graph is spatially filtered to remove noise. It is provided with a noise removing means for

A vehicle image processing apparatus according to a seventh aspect of the present invention is the vehicle image processing apparatus according to any one of the third to sixth aspects, in which the one-dimensional graph is subjected to time filtering to remove noise. It is provided with a noise removing means for

The vehicle image processing apparatus according to claim 8 of the present invention is the vehicle image processing apparatus according to claim 7, wherein the noise removing means changes the time constant of the time filter according to the vehicle speed or the like. The time constant changing means is provided.

[0018]

In the vehicle image processing apparatus according to the first aspect of the present invention, the edge detection means facilitates the edge detection processing and shortens the processing time.

In the vehicular image processing apparatus according to claim 2 of the present invention, the binarizing means binarizes the edge information, and the threshold varying means makes the threshold for binarizing variable so that it can be used for daytime or nighttime brightness. The edge information is detected appropriately according to the change in depth.

In the vehicle image processing apparatus according to the third aspect of the present invention, the edge projection means compresses the edge information by integrating the edge information in the vertical direction on the image to obtain a one-dimensional graph.

In the vehicle image processing apparatus according to the fourth aspect of the present invention, the horizontally long edge projection means detects horizontally long edge information, and the edge information is vertically integrated on the image to obtain a one-dimensional graph. Compress the edge information with.

In the vehicular image processing apparatus according to the fifth aspect of the present invention, the change amount detecting means detects an area in which the change amount of the one-dimensional graph obtained by the edge projecting means or the horizontally long edge projecting means is a predetermined value or more. The vehicle recognizing means detects the vehicle by recognizing it as a vehicle when the obtained variation is equal to or larger than a predetermined value.

In the vehicle image processing apparatus according to the sixth aspect of the present invention, the noise removing means removes background noise by performing spatial filter processing on the one-dimensional graph.

In the vehicular image processing apparatus according to the seventh aspect of the present invention, the noise removing means removes background noise by performing a time filtering process on the one-dimensional graph.

In the vehicular image processing device according to the eighth aspect of the present invention, the time constant changing means appropriately removes the background noise by changing the time constant of the time filter.

[0026]

【Example】

Example 1. An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of a vehicle image processing apparatus according to an embodiment of the present invention. The vehicle image processing apparatus according to the present embodiment is based on a CCD camera 11 which is a photographing means for photographing the periphery of the vehicle, an image memory 12 which stores an image output from the CCD camera, and data of the image memory 12. It is composed of a CPU 13 that performs image processing.

FIG. 2 is a diagram showing an outline of a processing flow in the CPU 13 in FIG. 1 together with each unit constituted by the CPU 13. First, the outline of the processing flow will be described with reference to FIG. In the CPU 13, the edge detection unit 21 detects a portion of the image input from the image memory 12 that has a brightness difference, and the binarization unit 22 binarizes it. Next, the edge information output by the binarization unit 22 is integrated in the vertical direction by the edge projection unit 23,
A one-dimensional graph is obtained, and the noise removing unit 24 smoothes the one-dimensional graph data. After that, the change amount detecting unit 25 detects the change amount of the projection data smoothed by the noise removing unit 24, and the vehicle recognizing unit 26 determines that the change amount is a vehicle.

Next, the details of each means of the processing flow shown in FIG. 2 will be described. FIG. 4 (a) shows a front image including the preceding vehicle photographed by the CCD camera 11, and FIG. 4 (b) performs processing by the edge detection means 21 and the binarization means 22 on this image. The resulting edge information is shown in the image state. Also, FIG.
(A) and (b) are the edge detection means 2 in this embodiment.
FIG. 5 is an explanatory diagram of the binarization means 22, and FIG.
(B) and (c) are the edge projection means 2 in this embodiment.
FIG. 3 is an explanatory diagram for explaining the operation of the noise removing unit 24, the change amount detecting unit 25, and the vehicle recognizing unit 26.

First, as shown in FIG. 3 (a), a certain input image 31 is displayed in the horizontal and vertical directions by n and m, respectively.
The image signal is shifted pixel by pixel, and the shifted image is compared with the image signal before shifting as 32. Image 3
FIG. 3B shows the edge information obtained by comparing 1 and 32 and binarizing the comparison. 31A,
32A is the image signal of the images 31 and 32, respectively, and 33 is the comparison result of the images 31 and 32. A threshold value Vth34 is set for this comparison result, and binarization is performed based on this threshold value.
A signal such as 5 is obtained as edge information. In this embodiment, the threshold value is set to be variable according to the iris control voltage of the CCD camera that adjusts the brightness of the entire image.

FIG. 4B shows an image obtained by executing the above-described processing in two dimensions. Figure 4
As shown in (b), after the edge information is detected and the binarization processing is performed, the shadow 42 and the guardrail 4 as well as the preceding vehicle 41 are displayed.
Those with contrast such as 3 remain in the image as edge information. An area 51 as shown in FIG. 5A is set on this image, and edge information in this area 51 is vertically integrated and projected on a one-dimensional graph. This is the edge projection unit 23, and the one-dimensional graph obtained by the edge projection unit 23 is shown by the graph 52 in FIG. 5B.
Is. Noise removing means 2 is applied to the one-dimensional graph 52.
4 removes background noise with a spatial filter. In this embodiment, as a spatial filter process for smoothing data, a process of taking an average of five adjacent projection data is
This is done for the entire projection data. As shown in FIG. 5B, the graph 52 made up of projection data that is strongly influenced by the background is smoothed like 53 by the process of taking an average of five adjacent data, and the shadow 42 and the guardrail 43 are obtained.
Background noise such as is removed.

The change detecting means 25 detects the change from the data smoothed by the noise removing means 24 described above. The graph 54 shown in FIG. 5C is a graph of this variation data. A threshold value 55 is set for the graph 54 of the variation data, and points 56 that are equal to or greater than the threshold value 55 are detected as vehicle candidate points. This is the vehicle recognition means 26.

In this embodiment, the edge projecting means for obtaining the one-dimensional graph by integrating the edge information of the image state in the vertical direction on the image is used, but among the edge information detected by the edge detecting means, Edges that are continuous in the horizontal direction (horizontally long edges) are detected by taking the logical product of the binarized edges of several pixels in the horizontal direction, and this horizontally long edge information is integrated vertically in the image. However, a laterally long edge projection means for obtaining a one-dimensional graph may be used. By using such a horizontally long edge projecting means, the edge of the vehicle is generally emphasized horizontally, so that it is possible to effectively remove the background such as the guardrail.

In this embodiment, the noise removing means 24
As the above, smoothing was performed by averaging adjacent data, but other spatial filter processing is also effective. In addition, the background has a large movement and a large change compared with the surrounding vehicles to be recognized. Therefore, instead of the above-mentioned processing, or by further performing a time filter processing on the above-mentioned processing, the time-series change of the projection data is performed. By smoothing, the background noise that is not stable in time series can be removed, and vehicle recognition can be performed more accurately. Further, in this case, the time constant varying means is a CPU.
13 is used, and the time constant of the time filter is made variable according to the speed of the host vehicle, noise can be accurately removed according to the vehicle speed.

[0034]

According to the vehicle image processing apparatus of the first aspect of the present invention, there are provided the photographing means for photographing the periphery of the vehicle and the edge detecting means for detecting the edge information from the image photographed by the photographing means. In the vehicle image processing device, the edge detecting means shifts the image obtained by the photographing means in the horizontal / vertical direction by a predetermined number of pixels, and compares the luminances of the images before and after the shift to determine the image state. Since the edge information is detected, the calculation becomes easier as compared with the edge extraction method using the filter, the hardware configuration scale can be reduced, and the processing speed is fast, and the vehicular image processing apparatus suitable for the vehicle can be obtained. There is an effect that can be.

According to the vehicle image processing apparatus of the second aspect of the present invention, in the vehicle image processing apparatus of the first aspect,
Since the edge detecting means includes the binarizing means for binarizing the detected edge information and the threshold varying means for varying the threshold when performing binarization by the binarizing means, during the daytime or Therefore, it is possible to obtain the vehicle image processing apparatus that can always obtain appropriate edge information from an image taken in an outdoor environment where the brightness changes, such as at night.

According to the vehicle image processing apparatus of claim 3 of the present invention, in the vehicle image processing apparatus of claim 1 or 2, the edge information detected by the edge detecting means is perpendicular to the image. To provide a vehicle image processing apparatus capable of compressing edge information, which is the amount of data to be handled, and accurately detecting a vehicle to be measured, because the vehicle has an edge projection unit that integrates in a direction to obtain a one-dimensional graph. There is an effect that can be.

According to the vehicle image processing apparatus of claim 4 of the present invention, in the vehicle image processing apparatus of claim 1 or 2, of the edge information of the image state detected by the edge detecting means, The horizontal edge projecting means for detecting the edge information indicating the horizontal edge and integrating the horizontal edge information in the vertical direction on the image to obtain a one-dimensional graph is used to calculate the edge information which is the amount of data to be handled. Thus, it is possible to obtain an image processing device for a vehicle that can be compressed and more accurately detect the vehicle that is the measurement target.

According to a fifth aspect of the present invention, there is provided a vehicle image processing device according to the third or fourth aspect, further comprising a change amount detecting means for detecting a change amount of the one-dimensional graph. When the change amount obtained by the change amount detecting means is equal to or more than a predetermined value, the vehicle recognizing means that determines that the vehicle is a vehicle is provided, so that the vehicle is separated from the background,
It is possible to obtain an image processing device for a vehicle that can be detected more accurately.

According to a sixth aspect of the present invention, there is provided a vehicle image processing apparatus according to any of the third to fifth aspects, wherein the one-dimensional graph is subjected to spatial filter processing to reduce noise. Since it is provided with the noise removing means for removing the background noise, it is possible to prevent the background noise from being erroneously detected as a vehicle, and it is possible to obtain a highly accurate vehicle image processing device.

According to a seventh aspect of the present invention, there is provided a vehicle image processing apparatus according to any one of the third to sixth aspects, wherein the one-dimensional graph is subjected to a temporal filtering process to reduce noise. Since it is equipped with a noise removal means that removes the background noise signal flowing on the image, the background can be accurately separated from the vehicle,
The vehicle image processing device capable of highly accurately recognizing the vehicle can be obtained.

According to the vehicle image processing apparatus of the eighth aspect of the present invention, in the vehicle image processing apparatus of the seventh aspect,
Since the noise removing means includes the time constant varying means for varying the time constant of the time filter according to the vehicle speed or the like, the effect of claim 7 can be obtained regardless of the vehicle speed. .

[Brief description of drawings]

FIG. 1 is a hardware configuration diagram of a vehicle image processing apparatus according to a first embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of a vehicle image processing apparatus according to the first embodiment.

FIG. 3 is an explanatory diagram of edge detection means according to the first embodiment.

FIG. 4 is a diagram showing an image of the front including a preceding vehicle taken by the vehicle image processing apparatus according to the first embodiment, and an output result of the edge detection unit.

FIG. 5 is an explanatory diagram of an edge projecting unit, a noise removing unit, a change detecting unit, and a vehicle recognizing unit of the vehicle image processing apparatus according to the first embodiment. .

FIG. 6 is an explanatory diagram of edge detection means in a conventional vehicle image processing apparatus.

FIG. 7 is a diagram showing a situation of outputting edge information of an image by a conventional vehicle image processing device.

FIG. 8 is a block diagram showing a configuration of a preceding vehicle recognition device using a conventional vehicle image processing device.

FIG. 9 is an explanatory diagram of a conventional preceding vehicle recognition means.

[Explanation of symbols]

 11 CCD Camera 12 Image Memory 13 CPU 21 Edge Detection Means 22 Binarization Means 23 Edge Projection Means 24 Noise Removal Means 25 Noise Detection Means 26 Vehicle Recognition Means 35 Binary Image Signals 41 Preceding Vehicle Image 42 Shadows 52 Projection Data 53 Data after noise removal 54 Change data 56 Vehicle candidate points

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI technical display location G08G 1/16 C 7531-3H

Claims (8)

[Claims] 1. An image processing apparatus for a vehicle, comprising: a photographing means for photographing the periphery of a vehicle; and an edge detecting means for detecting edge information from an image photographed by the photographing means, wherein the edge detecting means is the photographing means. An image processing apparatus for a vehicle, wherein the edge information is detected in an image state by shifting the image obtained by the means in a horizontal / vertical direction by a predetermined number of pixels and comparing the luminances of the images before and after the shifting. 2. The vehicle image processing device according to claim 1, wherein the edge detection means performs binarization by the binarization means for binarizing the detected edge information and the binarization means. An image processing apparatus for a vehicle, comprising: a threshold varying means for varying the threshold of. 3. The vehicle image processing apparatus according to claim 1, further comprising edge projection means for vertically integrating an edge information detected by the edge detection means on an image to obtain a one-dimensional graph. An image processing device for a vehicle characterized by the above. 4. The vehicle image processing apparatus according to claim 1 or 2, wherein edge information representing a horizontally long edge is detected from the edge information of the image state detected by the edge detecting means. An image processing apparatus for a vehicle, comprising a laterally long edge projection means for vertically integrating the laterally long edge information on an image to obtain a one-dimensional graph. 5. The vehicle image processing apparatus according to claim 3 or 4, wherein a change amount detecting means for detecting a change amount of the one-dimensional graph, and a change amount obtained by the change amount detecting means are predetermined values. A vehicle image processing device comprising: a vehicle recognition unit that determines that the vehicle is a vehicle when the above conditions are satisfied. 6. The vehicle image processing device according to claim 3, further comprising a noise removing unit that applies spatial filtering to the one-dimensional graph to remove noise. Image processing device. 7. The vehicle image processing device according to claim 3, further comprising noise removing means for removing noise by subjecting the one-dimensional graph to a time filtering process. Image processing device. 8. The vehicle image processing apparatus according to claim 7, wherein the noise removing means includes a time constant changing means for changing the time constant of the time filter according to the vehicle speed or the like. Vehicle image processing device.