基于传感器融合的轨道行人闯入识别技术研究

Research on Track Pedestrian Intrusion Recognition Technology Based on Sensor Fusion

单一传感器的识别算法在列车驾驶环境下鲁棒性较差,可靠性较低。3D激光雷达拥有极高的分辨率和较强的抗干扰能力,但由于激光雷达数据的特征单一,不如视觉检测的目标特征信息丰富。采用一种新颖的传感器融合方法,通过工业相机和3D激光雷达的融合为轨道提供增强的环境信息,能够更加准确地识别轨道中行人闯入的位置以及行人和列车的距离。

The identification algorithm of a single sensor has poor robustness and low reliability in the train driving environment. 3D lidar has extremely high resolution and strong anti-interference ability, but due to the single characteristics of the turbulent radar data, it is not as rich as the target feature information of visual detection. Therefore, a novel sensor fusion method is used in this paper to provide enhanced environmental information for the track through the fusion of industrial cameras and 3D lidar, which can more accurately identify the location of pedestrians intruding in the track and the distance between pedestrians and trains.