Target identification method and device, and computer readable storage medium

Abstract

The invention relates to a target identification method and device, and a computer readable storage medium. The method comprises the following steps: acquiring a driving environment image acquired by an image sensor and detection data obtained by radar scanning; performing space-time alignment on the driving environment image, the millimeter wave radar data and the point cloud data; generating a 3D region of interest according to the millimeter wave radar data after space-time alignment; respectively generating a first feature map corresponding to the point cloud data and a second feature map corresponding to the driving environment image by adopting a trained neural network; projecting the 3D region of interest to the first feature map and the second feature map to obtain a projection region of interest of the 3D region of interest; and fusing the projection region of interest, the first feature map and the second feature map, and identifying a target in the driving environment image. According to the scheme provided by the invention, the target can be identified with relatively high precision, and the real-time performance of identification is improved.

Description

technical field [0001] The present application relates to the technical field of intelligent driving, and in particular to a method, a device and a computer-readable storage medium for identifying a target. Background technique [0002] Object recognition is an important topic in the field of intelligent driving, because the accuracy and real-time of object recognition are related to driving safety. In related technologies, vision devices (such as monocular cameras, binocular cameras, or depth cameras, etc.) are mainly used to collect images of surrounding environment objects of smart cars, and then, combined with neural network models, target objects are identified with preset algorithms. However, the disadvantage of the above method is that it is greatly affected by environmental factors such as weather. When the environment is relatively harsh (for example, in heavy fog, heavy snow or strong wind), the recognition accuracy is poor due to the blurred image collected. In v...

AI Technical Summary

Problems solved by technology

However, since the amount of data collected by radar is usually large, this method brings a lot of consumption of computing resources, which leads to the disadvantage of poor real-time performance.

Images