Point cloud division and track matching method for millimeter wave radar target tracking

Abstract

The invention provides a point cloud division and track matching method for millimeter-wave radar target tracking, and belongs to the field of millimeter-wave radar signal processing. In order to improve the personnel counting accuracy of the millimeter-wave radar under the multi-target condition, the method comprises the following steps: transmitting an electromagnetic wave signal to a to-be-measured space range through the millimeter-wave radar, and receiving a signal reflected back in the to-be-measured space range through the millimeter-wave radar; processing the received signal to obtaina target trace point; aiming at the target plot, gathering the plot into different areas; matching the condensed trace point area with a known track according to conditions, distinguishing the motiontracks of different targets, obtaining the motion state of the personnel in the target area, and updating the counting result in the target area.

Description

technical field [0001] The invention relates to millimeter wave radar signal processing, in particular to a point cloud division and track matching method for millimeter wave radar target tracking. Background technique [0002] Personnel detection and quantity statistics are indispensable data in many existing scenarios, such as airports, shopping malls, chain stores, stations, museums, conference rooms, etc. Personnel detection and data estimation play an important role in safety management and improving the overall service level. important role. Common people detection technology mainly uses a single face recognition method, which is less accurate, such as when facing away from the camera. The access control system of RFID technology can monitor the entrance and exit of the area, but it needs to equip everyone with corresponding tags and deploy a large number of card reading devices, which is costly and poor in scalability and versatility. With the advancement of science...

AI Technical Summary

Problems solved by technology

When a moving target passes by another stationary target at a low speed, the points on the moving target are scattered, and a small number of dot traces will be matched to the stationary target because it is closer to the stationary target, resulting in tracking errors

However, some existing multi-objective clustering methods mainly perform clustering according to the position and speed of the dot traces. Since the number and distribution range of dot traces are closely related to the location and motion of the target, the commonly used clustering methods are easy to cause The number of clusters after clustering does not match the number of tracks, for example, when three moving targets move in an area where the track is relatively sparse, the track is only clustered into two clusters or the track is densely moving in an area with strong reflection The next 3 moving targets are divided into 4 or more clusters, resulting in some clusters being wrongly assigned to other targets and causing tracking errors. As the errors gradually accumulate, further target tracking and counting errors are caused.

Patent CN201811397869.9 A high-resolution millimeter-wave radar multi-target clustering method first eliminates the points with small snr values, and then clusters the dot traces according to the speed and position information of the dot traces. It does not use the snr peak dot traces, which is prone to clustering The number of clusters after the class is far and the number of moving objects is quite different

Patent CN201910515993.9 A radar multi-target clustering method and device first divides the dots into different areas according to the density of the dots, and then divides the dots again according to the difference between the three values ​​of position, speed and signal-to-noise ratio between the dots Divided into different areas, the number of moving targets cannot be determined more accurately

Images