Traffic target detection algorithm based on 4D radar and vision fusion

In the intelligent road traffic system, the sensor module is required to quickly distinguish each traffic target, but the traditional camera-based or millimeter-wave radar-based detection system is difficult to meet this requirement. Aiming at the problem that the limitation of a single sensor makes the longitudinal traffic target indistinct and the detection speed slow, a method for target detection is developed by combining 4D millimeter wave radar with a camera. The ROI of the target detection region of interest is constructed by using the characteristics of 4D millimeter wave radar with stronger perception ability and faster data processing, which reduces the computing power for the subsequent visual traffic target detection. The test results of the proposed detection algorithm in the experimental field show that the detection method is suitable for practical road traffic target detection applications.