Multi-laser radar decision level fusion method and device for pedestrian detection

Abstract

The invention relates to a multi-laser radar decision level fusion method and device for pedestrian detection, computer equipment and a storage medium. The method comprises the following steps of: performing pedestrian detection on point cloud data of a detection target acquired by each laser radar on an unmanned vehicle through a trained AdaBoost algorithm to obtain a pedestrian detection score of a single laser radar; conducting decision level fusion on the detection result of a radar pair formed by combining two laser radars on an unmanned vehicle through a Bayesian rule to obtain a pedestrian detection result of the radar pair, and then acquiring a final pedestrian detection result according to the pedestrian detection results of all the radar pairs in the multiple laser radars. According to the method, the single laser radar can firstly make decisions independently, then the decisions of the multiple laser radars are fused, data level fusion or feature level fusion of the multiplesensors is avoided, and laser radar data collection does not need to be completely synchronized, therefore the method has the advantages of small calculation amount and low requirement for the time sequence of original data of the laser radars.

Description

technical field [0001] The present application relates to the field of unmanned driving technology, in particular to a multi-lidar decision-level fusion method, device, computer equipment and storage medium for pedestrian detection. Background technique [0002] As an emerging technology combining artificial intelligence and automation technology, unmanned driving technology has gradually become an important driving force to promote the upgrading of the automobile industry and the penetration of robot technology into ordinary people's homes. At present, in order to improve the accuracy of detection and reduce the blind spot of detection, a large number of multi-line lidars are used in the research of unmanned vehicles to realize pedestrian detection. [0003] Traditional lidar data fusion methods include data-level and feature-level fusion. Data-level fusion is to directly fuse the raw data collected by sensors. Feature-level fusion is to first extract features (such as sha...

AI Technical Summary

Problems solved by technology

Due to less data processing before fusion, the traditional method retains the original information of the data to a large extent, but there are also defects such as large amount of data processing and poor real-time performance of the algorithm.

Images