摘要
Abstract
Agricultural machinery automatic navigation technology integrates global positioning system(GPS),laser ra-dar(LiDAR),cameras IMU and advanced algorithms enable agricultural machinery to independently carry out tasks such as positioning,path planning,obstacle detection and avoidance in the field,and achieve unmanned or semi-auto-matic driving.This paper summarizes the research progress and application status in this field at home and abroad,ana-lyzes the visual perception technology,feature extraction and recognition methods,navigation path planning strategies,and multi-sensor fusion positioning technology in the field environment,and analyzes the advantages and disadvantages of traditional image processing methods and deep learning algorithms in navigation target detection,operation boundary recognition,and path tracking through case comparison.It summarizes the technical routes and performance characteris-tics of typical systems.The research results show that high-precision visual perception,lightweight deep learning mod-els,edge computing processing,and full autonomous work collaboration will become the key development path.The re-search results aim to provide reference and inspiration for relevant researchers and engineering technicians,and promote the further development of intelligent and unmanned agricultural machinery.关键词
计算机视觉/农业机械/自动导航/综述/多传感器融合Key words
computer vision/agricultural machinery/autonomous navigation/review/multi-sensor fusion分类
农业科技
