Pedestrian posture recognition method and system based on radar and pattern recognition

Abstract

The invention relates to a pedestrian posture recognition method and system based radar and pattern recognition. The pedestrian posture recognition method comprises the steps of preprocessing an echosignal of a radar signal to obtain an output signal; performing suppression processing on a static object in the output signal; searching a distance unit in which a pedestrian is located in the suppression processed output signal; performing time-frequency analysis on an echo signal corresponding to the distance unit in which the pedestrian is located to obtain an echo signal time-frequency diagram; and recognizing the echo signal time-frequency diagram by using a convolutional neural network to obtain a human posture recognition result. According to the pedestrian posture recognition method based on radar and pattern recognition, the pedestrian posture recognition result is provided in real time through the analysis for radar echoes. The method is implemented based on the radar without being interfered by factors such as illumination conditions, weather and smoke, can work all day and all weather and has high engineering realizability, thereby being effectively performing high-accuracy recognition on the pedestrian posture. Meanwhile, the pedestrian posture recognition method is not easily affected by various environmental factors, and the system is good in robustness.

Description

technical field [0001] The invention relates to the technical field of radar signal processing and image recognition, in particular to a pedestrian gesture recognition method and system based on radar and pattern recognition. Background technique [0002] In the prior art, the posture recognition of pedestrians mostly uses an optical camera to acquire images, and then performs image recognition processing on the acquired images. In this way, due to the interference of environmental factors such as lighting conditions, weather, and smoke, especially at night, its recognition effect Greatly discounted, the recognition accuracy is not high, or even unrecognizable, and the stability is very poor. Therefore, this method can no longer meet occasions with high requirements such as 24 hours a day. Contents of the invention [0003] The technical problem to be solved by the present invention is to provide a pedestrian gesture recognition method and system based on radar and patter...

AI Technical Summary

Problems solved by technology

[0002] In the prior art, the posture recognition of pedestrians mostly uses an optical camera to acquire images, and then performs image recognition processing on the acquired images. In this way, due to the interference of environmental factors such as lighting conditions, weather, and smoke, especially at night, its recognition effect Greatly discounted, the recognition accuracy is not high, or even unrecognizable, and the stability is very poor

Therefore, this method can no longer meet the occasions with high requirements such as 24 hours a day

Images