Point cloud classification is quite challenging due to the influence of noise, occlusion, and the variety of types and sizes of objects. Currently, most methods mainly focus on subjectively designing and extracting features. However, the features rely on prior knowledge, and it is also difficult to accurately characterize the complex objects of point clouds. In this paper, we propose a concise multi-scale convolutional network (MSNet) for adaptive and robust point cloud classification. Both the local feature and global context are incorporated for this purpose. First, around each point, the spatial contexts of different sizes are partitioned as voxels of different scales. A voxel-based MSNet is then simultaneously applied at multiple scales to adaptively learn the discriminative local features. The class probability of a point cloud is predicted by fusing the features together across multiple scales. Finally, the predicted class probabilities of MSNet are optimized globally using the conditional random field (CRF) with a spatial consistency constraint. The proposed method was tested with data sets of mobile laser scanning (MLS), terrestrial laser scanning (TLS), and airborne laser scanning (ALS) point clouds. The experimental results show that the proposed method was able to achieve appreciable classification accuracies of 83.18%, 98.24%, and 97.02% on the MLS, TLS, and ALS data sets, respectively. The results also demonstrate that the proposed network has a strong generalization capability for classifying different kinds of point clouds under complex urban environments.

High-voltage and ultra-high-voltage overhead power lines are important to meet the electricity demand of our daily activities and productions. Due to the overgrowth of trees/vegetation within the corridor area, the distance between the power lines and its surroundings may break through the safety threshold, which could cause potential hazards such as discharge and fire. To ensure the safe and stable operation of the power lines, it is necessary to survey them regularly so that the potential hazards from the surroundings within the power line corridor could be investigated timely. This paper is motivated to quickly and accurately survey the power line corridor with the 3D point clouds. The main contributions of this paper include: (1) the spatial line clustering is proposed to accurately classify and complete the power line points, which can greatly overcome the sparsity and missing of LiDAR points within the complex power line corridor. (2) The contextual relationship between power lines and pylon is well investigated by the grid-based analysis, so that the suspension points of power lines on the pylon are well located. (3) The catenary plane-based simplification of 3D spatial distance calculation between power lines and ground objects facilitates the survey of the power line corridor. Experimental results show that the accuracy of safety distance surveying is 5 cm for power line corridors of all voltage levels. Compared to the ground-truth point-to-point calculation, the speed of surveying is enhanced thousands of times. It is promising to greatly improve both the accuracy and efficiency of surveying the potential hazards of power line corridor.

High-definition mapping of 3D lane lines has been widely needed for the highway documentation and intelligent navigation of autonomous systems. A mobile mapping system (MMS) captures both accurate 3D LiDAR point clouds and high-resolution images of lane markings at highway driving speeds, providing an abundant data source for combined lane mapping. This paper aims to map lanes with an MMS. The main contributions of this paper include the following: (1) an intensity correction method was introduced to eliminate the reflectivity inconsistency of road-surface LiDAR points; (2) a self-adaptive thresholding method was developed to extract lane markings from their complicated surroundings; and (3) a LiDAR-guided textural saliency analysis of MMS images was proposed to improve the robustness of lane mapping. The proposed method was tested with a dataset acquired in Wuhan, Hubei, China, which contained straight roads, curved roads, and a roundabout with various pavement markings and a complex roadside environment. The experimental results achieved a recall of 96.4%, a precision of 97.6%, and an F-score of 97.0%, demonstrating that the proposed method has strong mapping ability for various urban roads.

With the rapid increase of power supply demand, a large amount of stockpiles of coal have been formed during the process of coal excavation and transportation between the mines, ports, power plants and etc. Quantitative parameters, especially the volume of stockpile are important for the planning of coal production and consumption. Although laser scanning can collect the dense 3D coordinates of the stockpile surface for its quantification, stockpiles of coal have irregular shape, size, height, and base conditions, and may be overlapped with each other, which makes it hard to measure the different stockpiles automatically and accurately. This paper proposes an algorithm to extract and measure the stockpiles from the 3D point cloud data using the multi-scale directional curvature. Firstly, the second-order directional curvature analysis along multiple directions and at multiple scales is proposed to extract the distinctive ridge of crest in the point cloud of stockpiles. Then, starting with the crest points, a competitive growing strategy is proposed to accurately locate the points of slope in the stockpile. Finally, the stockpile’s volume is calculated by reconstructing the complete points of crest and slope with many meshes and triangular prisms through the subsurface fitting and surface reconstruction. Experiments on both the synthetic and real point cloud of stockpiles demonstrate that the proposed algorithm can extract the stockpiles with the average accuracy over 93.5% and measure the volume of stockpiles with the average precision over 93.7%. It is promising for automatically measuring the stockpiles like sand, soybean, etc., and facilitating the scientific storage and transportation management, as well as the maintenance of safety inventory during operation.

Multiple two-dimensional laser rangefinders (LRFs) are applied in many applications like mobile robotics, autonomous vehicles, and three-dimensional reconstruction. The extrinsic calibration between LRFs is the first step to perform data fusion and practical application. In this paper, we proposed a simple method to calibrate LRFs based on a corner composed of three mutually perpendicular planes. In contrast to other methods that require a special pattern or assistance from other sensors, the trihedron corner needed in this method is common in daily environments. In practice, we can adjust the position of the LRFs to observe the corner until the laser scanning plane intersects with three planes of the corner. Then, we formed a Perspective-Three-Point problem to solve the position and orientation of each LRF at the common corner coordinate system. The method was validated with synthetic and real experiments, showing better performance than existing methods.