Target Detection and Recognition of Radar spectrum Image Based on Yolov5 algorithm

The vehicle-mounted millimeter-wave radars have been widely mounted on modern smart cars, which greatly improves the accuracy of collecting surrounding environmental information, but driverless cars have higher requirements for safety [1]. Therefore, improving the performance of vehicle-mounted millimeter-wave radars is particularly important. Traditional vehicle-mounted millimeter-wave radars determine target information through the transmission and reception of echoes [2]. In the signal processing process, it is greatly affected by the outside world and it is difficult to maintain high accuracy. Therefore, a radar spectrum image detection method based on deep learning is proposed to quickly identify objects corresponding to the spectrum.First, the vehicle-mounted millimeter radar spectrum mat format data is extracted [3], make the dataset for training, then adjust and optimize the YOLOv5 model according to the situation of the data set, and identify the spectral image using the optimized a training model. In the experiment, 11635 pictures were divided into training set and test set at 9:1, and the average mAP of various targets reached 93.9%, thus detecting the radar echo spectrum based on deep learning and achieving good results. The proposed spectral target detection method [4] can be applied to vehicle-mounted millimeter-wave radar.

I. Introduction

During the first industrial revolution, in 1886, automobiles were manufactured for the first time. After more than a century of development in related fields, automobiles have become products that people cannot leave, providing great convenience for mankind. The annual demand for vehicles is increasing year by year to promote the mass production of vehicles. Technological innovations are always accompanied by potential safety hazards. Every year, countless lives die due to vehicle accidents, and the property losses are also huge. According to statistics from the Ministry of Public Security, as many as 16,600 people will die in traffic accidents in 2020.