Method and apparatus for recognizing object, device, vehicle and medium

1. A method for recognizing an object, comprising:

selecting a target camera matched with a target object from at least two cameras according to position information of the target object in a map and photographing parameters of the at least two cameras, wherein, the at least two cameras has at least two focal lengths;

controlling the target camera to photograph an image of the target object; and

obtaining the image photographed by the target camera, and performing recognition on the image to recognize the target object;

wherein, selecting the target camera matched with the target object from the at least two cameras according to position information of the target object in the map and photographing parameters of the at least two cameras comprises:

determining a projection position of the target object in an imaging plane of each camera based on the position information of the target object in the map and the photographing parameters of the at least two cameras, and determining the target camera, wherein the projection position of the target object locates in a photographing range of the imaging plane of the target camera;

wherein, determining the projection position of the target object in the imaging plane of each camera based on the position information of the target object in the map and the photographing parameters of the at least two cameras, and determining the target camera comprise:

selecting a camera as a current camera to be selected according to an ascending order of the focal lengths of the at least two cameras successively;

determining a projection position of the target object in an imaging plane of the current camera to be selected based on the position information of the target object in the map and a photographing parameter of the current camera to be selected; and

in response to determining that the projection position of the target object locates in the photographing range of the imaging plane of the current camera to be selected, determining the current camera to be selected as the target camera.

2. The method according to claim 1 , after determining the current camera to be selected as the target camera, further comprising: stopping determination operation of the target camera.

3. The method according to claim 1 , wherein, the photographing parameters comprises:

an internal parameter, comprising: the focal length, an optical center, and a distortion coefficient; and

an external parameter, comprising: a transformation matrix from a world coordinate system to a camera coordinate system.

4. The method according to claim 1 , before selecting the target camera matched with the target object from the at least two cameras, further comprising:

obtaining a position of a vehicle, wherein, the at least two cameras are installed on the vehicle; and

in response to determining that the vehicle locates at a crossroad according to the position of the vehicle, obtaining three-dimensional position information of a signal light at the crossroad from map data as the position information of the target object.

5. A device, comprising:

one or more processors, and

a storage device, configured to store one or more programs that, when executed by the one or more processors, cause the one or more processors to implement a method for recognizing an object, the method comprising:

selecting a target camera matched with a target object from at least two cameras according to position information of the target object in a map and photographing parameters of the at least two cameras, wherein, the at least two cameras has at least two focal lengths;

controlling the target camera to photograph an image of the target object; and

obtaining the image photographed by the target camera, and performing recognition on the image to recognize the target object;

wherein, selecting the target camera matched with the target object from the at least two cameras according to position information of the target object in the map and photographing parameters of the at least two cameras comprises:

determining a projection position of the target object in an imaging plane of each camera based on the position information of the target object in the map and the photographing parameters of the at least two cameras, and determining the target camera, wherein the projection position of the target object locates in a photographing range of the imaging plane of the target camera;

wherein, determining the projection position of the target object in the imaging plane of each camera based on the position information of the target object in the map and the photographing parameters of the at least two cameras, and determining the target camera comprise:

selecting a camera as a current camera to be selected according to an ascending order of the focal lengths of the at least two cameras successively;

determining a projection position of the target object in an imaging plane of the current camera to be selected based on the position information of the target object in the map and a photographing parameter of the current camera to be selected; and

in response to determining that the projection position of the target object locates in the photographing range of the imaging plane of the current camera to be selected, determining the current camera to be selected as the target camera.

6. The device according to claim 5 , wherein, after determining the current camera to be selected as the target camera, the method further comprises: stopping determination operation of the target camera.

7. The device according to claim 5 , wherein, the photographing parameters comprises:

an internal parameter, comprising: the focal length, an optical center, and a distortion coefficient; and

an external parameter, comprising: a transformation matrix from a world coordinate system to a camera coordinate system.

8. The device according to claim 5 , wherein before selecting the target camera matched with the target object from the at least two cameras, the method further comprises:

obtaining a position of a vehicle, wherein, the at least two cameras are installed on the vehicle; and

in response to determining that the vehicle locates at a crossroad according to the position of the vehicle, obtaining three-dimensional position information of a signal light at the crossroad from map data as the position information of the target object.

9. A vehicle, comprising: a vehicle body, a device, and at least two cameras disposed on the vehicle body, wherein, each of the at least two cameras employs a focal length, and the device comprises:

one or more processors, and

a storage device, configured to store one or more programs that, when executed by the one or more processors, cause the one or more processors to implement a method for recognizing an object, the method comprising:

selecting a target camera matched with a target object from at least two cameras according to position information of the target object in a map and photographing parameters of the at least two cameras, wherein, the at least two cameras has at least two focal lengths;

controlling the target camera to photograph an image of the target object; and

obtaining the image photographed by the target camera, and performing recognition on the image to recognize the target object;

wherein, selecting the target camera matched with the target object from the at least two cameras according to position information of the target object in the map and photographing parameters of the at least two cameras comprises:

determining a projection position of the target object in an imaging plane of each camera based on the position information of the target object in the map and the photographing parameters of the at least two cameras, and determining the target camera, wherein the projection position of the target object locates in a photographing range of the imaging plane of the target camera;

wherein, determining the projection position of the target object in the imaging plane of each camera based on the position information of the target object in the map and the photographing parameters of the at least two cameras and determining the target camera comprise:

selecting a camera as a current camera to be selected according to an ascending order of the focal lengths of the at least two cameras successively;

determining a projection position of the target object in an imaging plane of the current camera to be selected based on the position information of the target object in the map and a photographing parameter of the current camera to be selected; and

in response to determining that the projection position of the target object locates in the photographing range of the imaging plane of the current camera to be selected, determining the current camera to be selected as the target camera.

10. The vehicle according to claim 9 , wherein, after determining the current camera to be selected as the target camera, the method further comprises: stopping determination operation of the target camera.

11. The vehicle according to claim 9 , wherein, the photographing parameters comprises:

an internal parameter, comprising: the focal length, an optical center, and a distortion coefficient; and

an external parameter, comprising: a transformation matrix from a world coordinate system to a camera coordinate system.

12. The vehicle according to claim 9 , wherein before selecting the target camera matched with the target object from the at least two cameras, the method further comprises:

obtaining a position of a vehicle, wherein, the at least two cameras are installed on the vehicle; and

in response to determining that the vehicle locates at a crossroad according to the position of the vehicle, obtaining three-dimensional position information of a signal light at the crossroad from map data as the position information of the target object.