基于激光雷达多源数据融合的路障检测系统设计

Design of roadblock detection system based on lidar multi-source data fusion

摘要:

车辆行驶模式的改变会导致道路障碍观测盲区的出现，如何在满足车辆运行安全的条件下，对道路障碍进行快速检测已经成为了一项亟待解决的问题。基于此引入激光雷达多源数据融合理论，设计路障检测系统。利用多源以太网接口电路，提供激光图像传感器、路障特征提取分类器、TVPS150PBS图像解码器所需的应用电子量，再借助可编程逻辑器件，对待融合多源数据进行整合，完成基于激光雷达的路障检测系统硬件方案设计。在此基础上，规划雷达工作模式的识别环境，利用多源异构传感器设备，对检测角点特征进行提取处理，整合上述所有数据信息，完成雷达脉冲组序列的建模处理，实现基于激光雷达多源数据融合路障检测系统的顺利应用。实验结果表明，与DSP技术方案相比，激光雷达多源数据融合原理支持下路障检测系统的观测盲区更小，可在满足车辆运行安全的条件下，实现对道路障碍的快速检测。

Abstract:

The change of vehicle driving mode will lead to the occurrence of blind spot of road obstacle observation. How to quickly detect road obstacle under the condition of satisfying the safety of vehicle operation has become an urgent problem to be solved. Based on this, the multi-source data fusion theory of lidar is introduced to design the roadblock detection system. The multi-source Ethernet interface circuit is used to provide the application electronic quantity required by the laser image sensor, the roadblock feature extraction classifier and the TVPS150PBS image decoder. Then, with the help of the programmable logic device, the multi-source fusion data are integrated to complete the hardware design of the roadblock detection system based on lidar. On this basis, the identification environment of radar working mode is planned, and the detection corner features are extracted and processed by using multi-source heterogeneous sensor equipment. All the above data information is integrated to complete the modeling processing of radar pulse sequence, and the smooth application of the roadblock detection system based on lidar multi-source data fusion is realized. The experimental results show that, compared with DSP technology scheme, the roadblock detection system supported by Lidar multi-source data fusion principle has a smaller blind area of observation, and can realize the rapid detection of road obstacles under the condition of meeting the vehicle operation safety.