A real-time tracking method for fast moving obstacles based on laser sensor

Abstract

The invention provides a fast moving obstacle real-time tracking method based on a laser sensor. The fast moving obstacle real-time tracking method is characterized by comprising the following steps:single-line laser radars are mounted around a vehicle body, multi-line laser radars are mounted on the two sides of the vehicle body, a three-dimensional laser radar is mounted at the top of the vehicle body, and a coordinate system taking a laser radar as a center is established and is subjected to internal parameter calibration and external parameter calibration; a laser radar scanning mode is manually selected according to a road structure. The fast moving obstacle real-time tracking method based on the laser sensor, provided by the invention, is small in amount of calculation and good in tracking effect and can be widely used in the laser tracking field.

Description

technical field [0001] The invention belongs to the field of automatic driving, in particular to a laser sensor-based real-time tracking method for fast-moving obstacles. Background technique [0002] In today's electronic information society, target detection and tracking play a very important role in both military and civilian fields. In the field of military applications, accurate detection and tracking of enemy targets is the basis for precise strikes against each other, and it is also an important manifestation of military strength. In civilian applications, detection and tracking are also widely used in autonomous driving, drones, robots and other fields. [0003] The particle filter technology used in it is based on the Bayesian filter framework, through dynamic search for a set of random sample points with weights in the state space, and the posterior probability density p(X k |Z 1:k ) for approximate fitting, the weighted average value of the sample points is use...

AI Technical Summary

Problems solved by technology

As time goes on, more false positive results will be derived, which will greatly cause unnecessary calculation burden, affect the tracking effect, and even crash the system

[0005] Moreover, the detection and tracking performance of traditional particle filter algorithms depends largely on the number of particles, especially in real-world scenes, if the target is too small, the detection environment is noisy, obstacles are hidden, and objects appear quickly or In the case of disappearance, this will directly lead to an exponential increase in the amount of calculation and a decrease in real-time tracking performance, which often leads to a high false detection rate in tracking.

Images