Object detection device, driving assistance device, object detection method, and object detection program

1. An object detection device comprising:

a camera configured to image surroundings of a vehicle provided with the object detection device;

a control device including a processor configured to operate as:

a characteristic point extraction unit configured to extract characteristic points of another vehicle from an image captured by the camera;

a characteristic point grouping processing unit configured to group the characteristic points extracted by the characteristic point extraction unit and set a horizontal edge narrowing-down region based on the grouped characteristic point group;

a horizontal edge extraction unit configured to extract horizontal edges that are characteristic lines of approximately horizontal direction components in a search region of the image captured by the camera;

an overlap determination unit configured to extract the horizontal edges within the horizontal edge narrowing-down region; and

a detection object recognition unit configured to specify a specific horizontal edge satisfying a predetermined condition among the extracted horizontal edges within the horizontal edge narrowing-down region, set a region where the specific horizontal edge is a positional reference and is expanded toward all of the left, right, up, and down directions in the image as a recognition region, and recognize a detection object within the recognition region,

wherein the recognition region is set such that the specific horizontal edge is expanded by a first predetermined pixels toward the up direction, a left end of the specific horizontal edge is expanded by a second predetermined pixels toward the left direction, a right end of the specific horizontal edge is expanded by the second predetermined pixels toward the right direction, and the specific horizontal edge is expanded by a third predetermined pixels toward the down direction,

wherein the first predetermined pixels, the second predetermined pixels, and the third predetermined pixels are set based on a position of the specific horizontal edge on the image, and

wherein the horizontal edge narrowing-down region is set to be larger than a minimum rectangular region that contains the grouped characteristic point group.

2. The object detection device according to claim 1 ,

wherein the predetermined condition represents being positioned on a lowermost side among the horizontal edges extracted by the horizontal edge extraction unit.

3. The object detection device according to claim 1 ,

wherein the recognition region is set such that a component on an upper side of the specific horizontal edge is larger than a component on a lower side of the specific horizontal edge.

4. The object detection device according to claim 1 ,

wherein the detection object recognition unit performs enlargement or reduction correction of the horizontal edge narrowing-down region, in the case that the horizontal edge narrowing-down region differs from an estimated size of a detection object derived from its position in the image captured by the camera.

5. The object detection device according to claim 1 ,

wherein the detection object recognition unit performs enlargement correction of the horizontal edge narrowing-down region toward a center direction of the image captured by the camera, in the case that the horizontal edge narrowing-down region is set such that it is biased to either the left or the right from a center line with respect to the horizontal direction of the image captured by the camera.

6. The object detection device according to claim 1 ,

wherein in the case that the horizontal edge narrowing-down region is set such that it is biased to either the left or the right from a center line with respect to the horizontal direction of the image captured by the camera, the detection object recognition unit performs reduction correction of a larger of regions of the horizontal edge narrowing-down region when divided by the center line.

7. The object detection device according to claim 1 ,

wherein in the case that the horizontal edge narrowing-down region is set such that it is biased to either the left or the right from a center line with respect to the horizontal direction of the image captured by the camera, the detection object recognition unit performs enlargement correction of a smaller of regions of the horizontal edge narrowing-down region when divided by the center line.

8. A driving assistance device comprising;

an object detection device according to claim 1 , and

a driving assistance unit configured to perform driving assistance of the vehicle based on a detection result of the object detection device.

9. An object detection method comprising:

imaging surroundings of a vehicle with a camera;

extracting characteristic points of another vehicle from an image captured by the camera;

grouping the characteristic points into a characteristic point group and setting a horizontal edge narrowing-down region based on the characteristic point group;

extracting horizontal edges that are characteristic lines of approximately horizontal direction components in a search region of the image captured by the camera;

extracting the horizontal edges within the horizontal edge narrowing-down region;

specifying a specific horizontal edge satisfying a predetermined condition among the extracted horizontal edges within the horizontal edge narrowing-down region;

setting a region where the specific horizontal edge is a positional reference and is expanded toward all of the left, right, up, and down directions in the image as a recognition region; and

recognizing a detection object within the recognition region,

wherein the recognition region is set such that the specific horizontal edge is expanded by a first predetermined pixels toward the up direction, a left end of the specific horizontal edge is expanded by a second predetermined pixels toward the left direction, a right end of the specific horizontal edge is expanded by the second predetermined pixels toward the right direction, and the specific horizontal edge is expanded by a third predetermined pixels toward the down direction,

wherein the first predetermined pixels, the second predetermined pixels, and the third predetermined pixels are set based on a position of the specific horizontal edge on the image, and

wherein the horizontal edge narrowing-down region is set to be larger than a minimum rectangular region that contains the grouped characteristic point group.

10. An object detection system comprising:

a camera configured to image surroundings of a vehicle provided with the object detection device;

a control device including a processor and a non-transitory computer readable medium storing instructions for controlling the processor to operate as:

a characteristic point extraction unit configured to extract characteristic points of another vehicle from an image captured by the camera;

a characteristic point grouping processing unit configured to group the characteristic points extracted by the characteristic point extraction unit and set a horizontal edge narrowing-down region based on the grouped characteristic point group;

a horizontal edge extraction unit configured to extract horizontal edges that are characteristic lines of approximately horizontal direction components in a search region of the image captured by the camera;

an overlap determination unit configured to extract the horizontal edges within the horizontal edge narrowing-down region; and

a detection object recognition unit configured to specify a specific horizontal edge satisfying a predetermined condition among the extracted horizontal edges within the horizontal edge narrowing-down region, set a region where the specific horizontal edge is a positional reference and is expanded toward all of the left, right, up, and down directions in the image as a recognition region, and recognize a detection object within the recognition region,

wherein the recognition region is set such that the specific horizontal edge is expanded by a first predetermined pixels toward the up direction, a left end of the specific horizontal edge is expanded by a second predetermined pixels toward the left direction, a right end of the specific horizontal edge is expanded by the second predetermined pixels toward the right direction, and the specific horizontal edge is expanded by a third predetermined pixels toward the down direction,

wherein the first predetermined pixels, the second predetermined pixels, and the third predetermined pixels are set based on a position of the specific horizontal edge on the image, and

wherein the horizontal edge narrowing-down region is set to be larger than a minimum rectangular region that contains the grouped characteristic point group.