基于激光雷达感知的场景重构方法研究

Research on Ground Truth Generation Method Based on LiDAR Perception

针对现有环境感知模块的目标识别算法在实际应用中误识、漏识等问题,通过搭建基于激光雷达的测评系统,并根据场景重构的结果完成对其他类型感知模块的评测。该测评系统通过增加后处理算法,对激光雷达感知模块的数据进一步优化处理,有效提升环境重构范围及各类目标标注的准确率,提升其作为测评系统的评测能力和置信度。

To solve the problems of the existing environmental perception module in practical application,such as false and omission recognition,evaluation on other types of perception modules was carried out according to result of ground truth generation by establishing the LiDAR-based evaluation system.This evaluation system,by adding postprocessing algorithm,further optimizing LiDAR perception module data,effectively improves ground truth generation range and accuracy of various objects labeling,improves its evaluation ability and credibility as an evaluation system.