Vehicle-mounted radar echo noise reduction method and device, equipment and medium

• Main Authors: LIN FAN, HUANG FUKENG, ZHANG QIUZHEN
• Format: Patent
• Language: Chinese; English
• Published: 23.11.2021
• Subjects: ANALOGOUS ARRANGEMENTS USING OTHER WAVES; CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; HANDLING RECORD CARRIERS; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION ORRERADIATION OF RADIO WAVES; MEASURING; PHYSICS; PRESENTATION OF DATA; RADIO DIRECTION-FINDING; RADIO NAVIGATION; RECOGNITION OF DATA; RECORD CARRIERS; TESTING

The invention provides a vehicle-mounted radar echo noise reduction method and device, equipment and a medium. According to the invention, automatic noise reduction processing can be performed on echoes of a vehicle-mounted radar based on an echo noise reduction model constructed based on deep learning, and more data are learned with wider channels by fusing features of different convolutional layers, so that the learning performance higher than that of a traditional residual network is obtained under the same parameter background, automatic noise reduction processing can be carried out for different scenes, the processing efficiency is relatively high, the identification precision of the vehicle-mounted radar system is improved, and the probability of traffic accidents of vehicles is reduced. 本发明提供了一种车载雷达回波降噪方法、装置、设备及介质，能够基于深度学习构建的回波降噪模型对车载雷达的回波进行自动降噪处理，通过对不同卷积层的特征进行融合，以更宽的通道数学习更多的数据，使得在同样的参数背景下，获得比传统残差网络更强大的学习性能，并且能够针对不同场景进行自动降噪处理，处理效率较高，从而提高了车载雷达系统的识别精度，降低了车辆发生交通事故的概率。