Shi Tiandong, Zhong Deyun, Bi Lin
School of Resources and Safety Engineering, Central South University, Changsha 410083, China.
Sensors (Basel). 2021 May 19;21(10):3548. doi: 10.3390/s21103548.
In transportation at open-pit mines, rocks dropped as a mining truck is driven will wear out the tires of the vehicle, thus increasing the mining cost. In the case of autonomous vehicles, the vehicle must automatically detect rocks on the transportation roads during the driving process. This will be a new challenge: rough road, rocks of small size and irregular shape, long detection distance, etc. This paper presents a detection method based on light detection and ranging (lidar). It includes two stages: (1) using the modified cloth simulation method to filter out the ground points; (2) using the regional growth method based on grid division to cluster non-ground points. Experimental results show that the method can detect rocks with a size of 20-30 cm at a distance of 40 m in front of the vehicle, and it takes only 0.3 s on an ordinary personal computer (PC). This method is easy to understand, and it has fewer parameters to be adjusted. Therefore, it is a better method for detecting small, irregular obstacles on a low-speed, unstructured and rough road.
在露天矿的运输过程中，采矿卡车行驶时掉落的岩石会磨损车辆轮胎，从而增加采矿成本。对于自动驾驶车辆而言，车辆在行驶过程中必须自动检测运输道路上的岩石。这将是一个新的挑战：道路崎岖、岩石尺寸小且形状不规则、检测距离长等。本文提出了一种基于激光雷达的检测方法。它包括两个阶段：（1）使用改进的布料模拟方法滤除地面点；（2）使用基于网格划分的区域生长方法对非地面点进行聚类。实验结果表明，该方法能够在车辆前方40米处检测到尺寸为20 - 30厘米的岩石，在普通个人计算机（PC）上仅需0.3秒。该方法易于理解，需要调整的参数较少。因此，它是一种在低速、非结构化且崎岖道路上检测小型不规则障碍物的较好方法。
