A fast and optimal pathfinder using airborne LiDAR data

Determining the optimal path between two points in a 3D point cloud is a problem that have been addressed in many different situations: from road planning and escape routes determination, to network routing and facility layout. This problem is addressed using different input information, being 3D point clouds one of the most valuables. Its main utility is to save costs, whatever the field of application is. In this paper, we present a fast algorithm to determine the least cost path in an Airborne Laser Scanning point cloud. In some situations, like finding escape routes for instance, computing the solution in a very short time is crucial, and there are not many works developed in this theme. State of the art methods are mainly based on a digital terrain model (DTM) for calculating these routes, and these methods do not reflect well the topography along the edges of the graph. Also, the use of a DTM leads to a significant loss of both information and precision when calculating the characteristics of possible routes between two points. In this paper, a new method that does not require the use of a DTM and is suitable for airborne point clouds, whether they are classified or not, is proposed. The problem is modeled by defining a graph using the information given by a segmentation and a Voronoi Tessellation of the point cloud. The performance tests show that the algorithm is able to compute the optimal path between two points by processing up to 678,820 points per second in a point cloud of 40,000,000 points and 16 km(2) of extension.

Automated summary

This paper presents a fast algorithm for determining the least cost path in an Airborne Laser Scanning point cloud, which avoids the loss of information and precision caused by using a digital terrain model (DTM) in traditional methods.