TRACK PREDICTION METHOD AND DEVICE FOR OBSTACLE AT JUNCTION
A track prediction method and device for an obstacle at a junction are provided. The method includes: acquiring environment information of a junction to be passed by a vehicle, and acquiring information on a visible obstacle in a sensible range of the vehicle, wherein the environment information comprises road information and information on a junction obstacle located in an area of the junction; combining the information on the junction obstacle with the information on the visible obstacle, and selecting information on a blind-zone obstacle in a blind zone of the vehicle at the junction; and predicting a moving track of an obstacle corresponding to the information on the blind-zone obstacle, according to the road information.
1 . A track prediction method for an obstacle at a junction, comprising:
acquiring environment information of a junction to be passed by a vehicle, and acquiring information on a visible obstacle in a sensible range of the vehicle, wherein the environment information comprises road information and information on a junction obstacle located in an area of the junction;
combining the information on the junction obstacle with the information on the visible obstacle, and selecting information on a blind-zone obstacle in a blind zone of the vehicle at the junction; and
predicting a moving track of an obstacle corresponding to the information on the blind-zone obstacle, according to the road information.
2 . The track prediction method of claim 1 , wherein acquiring environment information of a junction to be passed by a vehicle comprises:
when a distance between the vehicle and the junction reaches a preset distance, receiving the environment information acquired by an acquisition device at the junction.
3 . The track prediction method of claim 1 , wherein combining the information on the junction obstacle with the information on the visible obstacle, and selecting information on a blind-zone obstacle in a blind zone of the vehicle at the junction comprises:
matching the information on the junction obstacle with the information on the visible obstacle;
determining whether the information on the junction obstacle and the information on the visible obstacle comprise identical information; and
if the information on the junction obstacle and the information on the visible obstacle comprise information on a same obstacle, obtaining the information on the blind-zone obstacle by removing the identical information from the information on the junction obstacle.
4 . The track prediction method of claim 3 , wherein determining whether the information on the junction obstacle and the information on the visible obstacle comprise identical information comprises:
acquiring historical frame data of a first obstacle located at the junction, based on the information on the junction obstacle;
acquiring historical frame data of a second obstacle in a sensible range of the vehicle, based on the information on the visible obstacle;
performing feature matching to the historical frame data of the first obstacle and the historical frame data of the second obstacle by using a preset model; and
when a matching result is greater than a preset threshold, determining that information corresponding to the first obstacle and information corresponding to the second obstacle are identical.
5 . The track prediction method of claim 1 , further comprising:
determining whether the information on the junction obstacle comprises information on the vehicle; and
if the information on the junction obstacle comprises the information on the vehicle, removing the information on the vehicle from the information on the junction obstacle.
6 . The track prediction method of claim 1 , wherein predicting a moving track of an obstacle corresponding to the information on the blind-zone obstacle according to the road information comprises:
acquiring historical frame data of the blind-zone obstacle based on the information on the obstacle; and
predicting a moving track of the blind-zone obstacle at the junction, according to junction environment information and signal light state information in the road information in combination with the historical frame data of the obstacle.
7 . A track prediction device for an obstacle at a junction, comprising:
one or more processors; and
a storage device configured for storing one or more programs, wherein
the one or more programs are executed by the one or more processors to enable the one or more processors to:
acquire environment information of a junction to be passed by a vehicle, and acquire information on a visible obstacle in a sensible range of the vehicle, wherein the environment information comprises road information and information on a junction obstacle located in an area of the junction;
combine the information on the junction obstacle with the information on the visible obstacle, and select information on a blind-zone obstacle in a blind zone of the vehicle at the junction; and
predict a moving track of an obstacle corresponding to the information on the blind-zone obstacle, according to the road information.
8 . The track prediction device of claim 7 , wherein the one or more programs are executed by the one or more processors to enable the one or more processors further to:
when a distance between the vehicle and the junction reaches a preset distance, receive the environment information acquired by an acquisition device at the junction.
9 . The track prediction device of claim 7 , wherein the one or more programs are executed by the one or more processors to enable the one or more processors further to:
match the information on the junction obstacle with the information on the visible obstacle; and
determine whether the information on the junction obstacle and the information on the visible obstacle comprise identical information; and if the information on the junction obstacle and the information on the visible obstacle comprise information on a same obstacle, obtain the information on the blind-zone obstacle by removing the identical information from the information on the junction obstacle.
10 . The track prediction device of claim 9 , wherein the one or more programs are executed by the one or more processors to enable the one or more processors further to:
acquire historical frame data of a first obstacle located at the junction, based on the information on the junction obstacle;
acquire historical frame data of a second obstacle in a sensible range of the vehicle, based on the information on the visible obstacle;
perform feature matching to the historical frame data of the first obstacle and the historical frame data of the second obstacle by using a preset model; and
when a matching result is greater than a preset threshold, determine that information corresponding to the first obstacle and information corresponding to the second obstacle are identical.
11 . The track prediction device of claim 7 , the one or more programs are executed by the one or more processors to enable the one or more processors further to:
determine whether the information on the junction obstacle comprises information on the vehicle; and if the information on the junction obstacle comprises the information on the vehicle, remove the information on the vehicle from the information on the junction obstacle.
12 . The track prediction device of claim 7 , wherein the one or more programs are executed by the one or more processors to enable the one or more processors further to:
acquire historical frame data of the blind-zone obstacle based on the information on the obstacle; and
predict a moving track of the blind-zone obstacle at the junction, according to junction environment information and signal light state information in the road information in combination with the historical frame data of the obstacle.
13 . A non-volatile computer-readable storage medium, storing a computer program executable instructions stored thereon, that when executed by a processor cause the processor to perform operations comprising:
acquiring environment information of a junction to be passed by a vehicle, and acquiring information on a visible obstacle in a sensible range of the vehicle, wherein the environment information comprises road information and information on a junction obstacle located in an area of the junction;
combining the information on the junction obstacle with the information on the visible obstacle, and selecting information on a blind-zone obstacle in a blind zone of the vehicle at the junction; and
predicting a moving track of an obstacle corresponding to the information on the blind-zone obstacle, according to the road information.
14 . The non-volatile computer-readable storage medium of claim 13 , wherein the computer executable instructions, when executed by a processor, cause the processor to perform further operations comprising:
when a distance between the vehicle and the junction reaches a preset distance, receiving the environment information acquired by an acquisition device at the junction.
15 . The non-volatile computer-readable storage medium of claim 13 , wherein the computer executable instructions, when executed by a processor, cause the processor to perform further operations comprising:
matching the information on the junction obstacle with the information on the visible obstacle;
determining whether the information on the junction obstacle and the information on the visible obstacle comprise identical information; and
if the information on the junction obstacle and the information on the visible obstacle comprise information on a same obstacle, obtaining the information on the blind-zone obstacle by removing the identical information from the information on the junction obstacle.
16 . The non-volatile computer-readable storage medium of claim 15 , wherein the computer executable instructions, when executed by a processor, cause the processor to perform further operations comprising:
acquiring historical frame data of a first obstacle located at the junction, based on the information on the junction obstacle;
acquiring historical frame data of a second obstacle in a sensible range of the vehicle, based on the information on the visible obstacle;
performing feature matching to the historical frame data of the first obstacle and the historical frame data of the second obstacle by using a preset model; and
when a matching result is greater than a preset threshold, determining that information corresponding to the first obstacle and information corresponding to the second obstacle are identical.
17 . The non-volatile computer-readable storage medium of claim 13 , wherein the computer executable instructions, when executed by a processor, cause the processor to perform further operations comprising:
determining whether the information on the junction obstacle comprises information on the vehicle; and
if the information on the junction obstacle comprises the information on the vehicle, removing the information on the vehicle from the information on the junction obstacle.
18 . The non-volatile computer-readable storage medium of claim 13 , wherein the computer executable instructions, when executed by a processor, cause the processor to perform further operations comprising:
acquiring historical frame data of the blind-zone obstacle based on the information on the obstacle; and
predicting a moving track of the blind-zone obstacle at the junction, according to junction environment information and signal light state information in the road information in combination with the historical frame data of the obstacle.