Barrier identification method and apparatus, computer device and readable medium

Abstract

The invention provides a barrier identification method and apparatus, a computer device and a readable medium. The method comprises the steps of obtaining information of a to-be-identified barrier of N+1 continuous frames around a current vehicle scanned by a laser radar; according to the information of the to-be-identified barrier in each of the first N frames, obtaining first point cloud projection drawings, on a horizontal plane, of a point cloud layer with at least two heights, a first reflection information projection drawing, on the horizontal plane, of the to-be-identified barrier, and a first duty ratio projection drawing, on the horizontal plane of the to-be-identified barrier in each frame; and according to a pre-trained classifier model, a pre-obtained first barrier type graph, on the horizontal plane, of the to-be-identified barrier in each of the first N frames, and the at least two first point cloud projection drawings, the first reflection information projection drawing and the first duty ratio projection drawing of each of the first N frames, predicting a first barrier type graph of the (N+1)th frame. According to the technical scheme, the identification accuracy and efficiency of the to-be-identified barrier can be effectively improved.

Description

[0001] 【Technical field】 [0002] The present invention relates to the technical field of automatic driving, in particular to an obstacle identification method and device, computer equipment and a readable medium. [0003] 【Background technique】 [0004] In the existing automatic driving technology, the output information of the obstacle to be recognized will be used as the input of control and planning information. Therefore, accurate and fast recognition of the obstacle to be recognized is a very critical technology. [0005] In the prior art, a camera and a laser radar are usually used to identify obstacles to be identified. Among them, the camera solution can be applied in scenes with sufficient light and relatively stable environment. However, in the case of bad weather and chaotic road environment, the vision of the camera solution has not been stable enough, resulting in inaccurate information collected about obstacles to be identified. Although lidar is very expensive...

AI Technical Summary

Problems solved by technology

[0006] However, in the prior art, when the obstacle to be identified is detected based on the point cloud of the obstacle to be identified in a single frame scanned by the lidar, all obstacles to be identified are static, for example, it is easy to identify pedestrians as background Columnar objects in the road, etc., resulting in poor recognition accuracy and low recognition efficiency of obstacles to be recognized in the road

Images