Authors

Abstract

This paper proposes a path planning method for phenotyping vehicles using UAV imagery and an enhanced A* algorithm. Orthophotos and Digital Surface Model (DSM) are generated from UAV images. Regions of Interest are extracted and rectified, followed by plot boundary segmentation using an energy function. Navigable areas are derived from DSM, and a roll angle model is developed to determine traversal costs. These costs are integrated into an enhanced A* algorithm incorporating jump point search and diagonal distance heuristics. Simulation and field trials show that the method significantly reduces travel distance and time while avoiding uneven terrain.

Abstract in Chinese

本文提出一种基于无人机影像与改进A*算法的表型车辆路径规划方法。通过无人机影像生成正射影像与数字表面模型（DSM），提取并校正感兴趣区域后，利用能量函数分割小区边界。基于DSM提取可通行区域，并建立侧倾角模型以确定通行成本。将该成本融入引入跳点搜索与对角距离启发式的改进A*算法中进行路径规划。仿真与田间试验结果表明，该方法显著缩短了行驶距离与时间，并能有效避开不平坦地形。