JPH10214395A - Lane change detector - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a high reliability against the change of environments such as a weather and a quickness for real-time processing. SOLUTION: An image pickup means 1 which picks up the image of a road, an area division means 3 which divides the area of the input image picked up by the image pickup means 1 in accordance with lanes on the surface of the road, an edge extraction means 5 which extracts edges of the image in each of lane areas divided by the area division means 3, and an edge number count means which counts the number of edges extracted for each lane area by the edge extraction means 5 are provided. Further, the edge number count means 7 is provided with a running lane discrimination means 9 which discriminates it at the time of the change of the counted number in two or more areas that the running lane of the vehicle is changed on the surface of the road.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a lane change detecting device, and more particularly to a lane change detecting device which is installed on a road and detects a lane change of a vehicle on the road surface by image processing.

[0002]

2. Description of the Related Art Conventionally, lane changes of a vehicle are recognized by image processing by lane recognition processing, vehicle recognition processing, and the like.
The lane change recognition process was required. As a method of performing lane recognition, for example, a method of individually recognizing left and right white lines by edge extraction processing is disclosed in Japanese Patent Laid-Open No. 6-341182.
No. 1 discloses this.

[0003] Recognition of a vehicle has been performed, for example, by taking a difference from a background image. The lane change detection process is performed by tracing the trajectory of the cut-out vehicle.

[0004]

The first problem is that, in the conventional lane change detection by the background subtraction and the tracing process, the cutout of the vehicle due to the background difference is vulnerable to changes in weather conditions and the like. And so on to be worse.

[0005] The reason is that the brightness of the background itself changes due to a change in the weather, so that the brightness of the entire screen on which image processing is performed changes, making it impossible to cut out the vehicle.

A second problem is that real-time processing is difficult.

[0007] The reason is that the tracing process of a vehicle has a very large processing amount, so that it is difficult to develop hardware for high-speed processing.

[0008]

An object of the present invention is to improve the disadvantages of the prior art and, in particular, to provide a lane change detecting device capable of obtaining high reliability against environmental changes such as weather and high speed capable of real-time processing. The purpose is to provide.

[0009]

Therefore, according to the present invention, an image pickup means for picking up an image of a road, an area dividing means for dividing an area of an input image picked up by the image pickup means in accordance with the lane of the road, Edge extracting means for extracting an edge of an image for each lane area divided by the area dividing means; and edge number counting means for counting the number of edges extracted for each lane area by the edge extracting means. I have. Moreover, the edge number counting means is provided with traveling lane determining means for judging that the traveling lane of the vehicle has changed on the road when the count number for each area changes in two or more areas. ing. This aims to achieve the above-mentioned object.

The edge extracting means extracts an edge of the image for each lane area divided by the area dividing means. When a vehicle is present on the road, the number of edges increases, and when the edge number counting means counts the number of edges for each region, information corresponding to the presence of the vehicle in each lane region is obtained. Furthermore, the traveling lane determining means determines that the traveling lane of the vehicle has changed on the road when the count number for each area changes in two or more areas. This is because when the vehicle changes course and departs from one lane area, the number of edges in the one lane area decreases, and furthermore, when the vehicle whose lane changed changes enters the other lane area, The fact that the number of edges in the other lane area increases is used.
At this time, when the number of edges changes in only one lane area, it can be determined whether the vehicle traveling straight in the lane area has entered or exited the imaging range. It is determined that the lane change of the vehicle has occurred when the count number changes in the area of 2 or more.

[0011]

Next, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing the configuration of the lane change detection device according to the present embodiment. The lane change detection device includes an image pickup unit 1 for picking up an image of a road, an area dividing unit 3 for dividing an area of an input image picked up by the image pickup unit 1 in accordance with a lane on the road surface, An edge extraction unit 5 for extracting an edge of an image for each of the extracted lane regions, and an edge number counting unit 7 for counting the number of edges extracted for each lane region by the edge extraction unit 5 are provided.
In addition, the edge number counting means 7 is provided with a traveling lane determining means 9 for judging that the traveling lane of the vehicle has changed on the road surface when the count number for each area changes in two or more areas.

In some embodiments, when the number of edges in the one lane area increases in the other lane area, the traveling lane determining means 9 determines whether or not the number of edges in the other lane area increases. It is preferable to provide a function for determining the number of edges to be increased / decreased to determine that the vehicle has changed the traveling lane to the lane.
By this edge number increase / decrease number determination function, it is possible to accurately determine from which lane the vehicle has changed lanes.

The image pickup means 1 is fixed on a road or mounted on a vehicle. When the imaging unit 1 is fixed on the road, the lane area within the imaging range is determined in advance, so the area dividing unit 3 defines an area for the imaging range for each lane in advance. On the other hand, when the imaging unit is mounted on a vehicle, the area dividing unit 3 performs white line recognition processing and the like to divide the imaging range for each lane.

The edge extracting means 5 extracts, for each lane area, an edge whose density changes within the area. In one embodiment, the edge is extracted after the edge is emphasized by the Sobel filter processing. The direction in which edges are emphasized is not limited to the horizontal direction, and edges in all directions are extracted. Since the purpose of this edge extraction is to detect a vehicle, an edge having a certain length or less may be discarded as noise.

The edge number counting means 7 counts the number of edges for each lane area by performing processing such as labeling the edges extracted by the edge extracting means 5 and counting the number of labels.

The traveling lane determining means 9 determines, based on the number of edges counted by the edge number counting means 7, a change in the traveling lane of the vehicle existing within the imaging range.
When the number of edges changes between the two regions, it is determined that a lane change has occurred between the regions. Therefore, the lane change can be accurately performed without erroneous detection due to a change in the number of edges due to a vehicle entering the imaging range. Changes can be detected.

[0017]

Next, an embodiment of the present invention will be described. In the present embodiment, the edge extracting means 5 includes a Sobel filter processing unit 6 for enhancing the edge of the image for each lane area. The edge extracting means 5 and the edge number counting means 7
Is processed by the parallel processor 8. That is,
The edge extraction means 5 has a parallel processing function of processing the emphasis and extraction of the edge of the image for each lane region in parallel for each image of each vehicle region.

FIG. 3 is an explanatory view showing an example in which an image taken according to the present embodiment is schematically displayed. In the example shown in FIG. 3, the imaging means 1 is fixedly installed on a highway. In response to this, the area dividing means 3 sets the area in advance based on the image captured by the image capturing means 1. As shown in FIG. 3, in the present embodiment, since there are three lanes, the area is divided into areas A to C.

In the present embodiment, a sobel filter processing unit capable of performing high-speed processing using a parallel processor is provided. Since the sobel filter processing is hardly affected by environmental changes such as weather, detection accuracy is improved. Further, in order to increase the speed of the Sobel filter processing and the processing of counting the processing, real-time processing can be performed using hardware constituted by a parallel processor dedicated to image processing.

Further, since the algorithm for detecting the lane change is based on the history of the edge amount change, the amount of processing is small and the lane change can be detected at high speed.

Next, the operation of this embodiment will be described. First, as shown in FIG. 3, the road surface image is divided into a traveling lane (hereinafter A), a center line (hereinafter B), and an overtaking lane (hereinafter C). Next, image processing is performed for each of the A section, the B section, and the C section, and the number of edges is detected. Further, the number of edges is counted for each of the A section, the B section, and the C section.

Finally, a change in the number of edges is detected for each of the portions A, B, and C, and a lane change is determined from each edge change of the portions A, B, and C.

In this method, vehicles are detected by utilizing the fact that the number of edges on a road surface is small and the amount of edges increases when vehicles are present. The lane change determination method based on the change in the number of edges of the A section, the B section, and the C section is as follows. Considering the case where the lane is changed from the traveling lane to the overtaking lane, the number of edges of the A section increases because a vehicle is present in the A section. . Next, since the vehicle moves from the part A to the part B, the number of edges of the part A decreases,
The number of edges of the part increases. As described above, when it is detected that the number of edges of the portion A decreases and the number of edges of the portion B increases, it is determined that the lane has been changed.

Similarly, when the lane is changed from the overtaking lane to the traveling lane, the number of edges in the portion C decreases, and the number of edges in the portion B increases. When such a change is detected, it is determined that the lane has been changed.

As described above, according to the present embodiment, since the Sobel filter processing is used for vehicle detection, the edges are relatively stable even if there is a luminance change, so that the accuracy of vehicle detection is high. Therefore, detection accuracy is high with respect to environmental changes such as weather.

Since a dedicated parallel processor is used for the hardware, the Sobel filter processing and the edge count processing can be performed at high speed.

Since the algorithm for detecting the lane change is based on the detection of a change in the number of edges instead of the tracing process of the vehicle, the lane change can be detected with a small amount of processing, so that the lane change can be detected at high speed.

[0028]

According to the present invention, the edge extracting means extracts the edge of the image for each lane area divided by the area dividing means, and the traveling lane judging means operates. When the count of the number of edges in each area changes in an area of 2 or more, it is determined that the traveling lane of the vehicle has changed on the road. It is possible to detect the lane change of the vehicle caused by the, particularly, when the vehicle changes course and deviates from one lane region, the number of edges in the one lane region decreases, and further, the traveling lane When the changed vehicle enters the other lane area, the number of edges in the other lane area increases, so that the lane change can be detected with high accuracy, and the count number for each area changes in the area of 2 or more. When Conventionally, it is possible to properly detect only a lane change without erroneous detection when a vehicle traveling straight in the lane area enters or exits the imaging range to determine that a lane change of the vehicle has occurred. An excellent lane change detection device can be provided.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an embodiment according to the present invention.

FIG. 2 is a block diagram showing a configuration of one embodiment according to the present invention.

FIG. 3 is an explanatory diagram showing an example of area division according to the embodiment.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Imaging means 3 Area division means 5 Edge extraction means 7 Edge number counting means 9 Traveling lane judgment means

Claims (5)

[Claims] 1. An image pickup means for picking up an image of a road surface, an area dividing means for dividing an area of an input image picked up by the image pickup means in accordance with a lane of the road surface, and each lane area divided by the area dividing means. Edge extracting means for extracting the edges of the image, and edge number counting means for counting the number of edges extracted for each lane region by the edge extracting means, respectively. A lane change detecting device provided with a traveling lane determining means for judging that the traveling lane of the vehicle has changed on the road surface when the count number of the vehicle changes in an area of 2 or more. 2. The vehicle according to claim 1, wherein said traveling lane determining means determines that said vehicle travels from said one lane to the other lane when the number of edges substantially equal to the decrease in the number of edges in one lane region increases in the other lane region. 2. The lane change detecting device according to claim 1, further comprising a judgment function for differentiating the number of edges to determine that the lane has been changed. 3. The lane change detecting device according to claim 1, wherein said edge extracting means includes a Sobel filter processing unit for enhancing an edge of the image for each lane area. 4. The apparatus according to claim 3, wherein said edge extracting means has a parallel processing function of processing emphasis and extraction of an edge of the image for each lane area in parallel for each image of each vehicle area. The lane change detection device according to the above. 5. The apparatus according to claim 1, wherein the image pickup unit is fixedly installed on a highway, and the region dividing unit sets the region in advance based on an image picked up by the image pickup unit. The lane change detection device according to the above.