摘要
传统的农业耕作方式,尤其是除草作业,已难以满足现代农业对高效、低成本和可持续发展的需求。本文结合了计算机视觉、深度学习和路径规划技术,提出一种基于人工智能(AI)加持的智能除草视觉导航方法,利用摄像头捕捉农田图像,通过图像处理和目标检测算法识别杂草与作物的差异,进而精准定位杂草的位置。基于识别结果,算法能够计算出最优除草路径,在不同环境下的识别准确率达到96%,机器人能够在每小时完成约200 m~2的除草任务,系统的避障成功率达到98%。在实际试验中,机器人能够成功避开作物、石块、路面起伏,同时能够根据实时反馈动态调整路径,在复杂地形下高效导航。研究结果可为智能农业领域的除草技术提供新的思路和解决方案,为推动农业智能化、精细化管理提供技术支持。
Traditional agricultural farming methods,especially weeding operations,have been difficult to meet the needs of modern agriculture for high efficiency,low cost and sustainable development.This paper combines computer vision,deep learning and path planning techniques to propose an intelligent weeding visual navigation method based on artificial intelligence(AI)augmentation,which uses a camera to capture images of the farmland,and recognises the differences between weeds and crops through image processing and target detection algorithms,and then accurately locates the position of the weeds.Based on the recognition results,the algorithm is able to calculate the optimal weeding path,with a recognition accuracy of more than 96%in different environments,the robot is able to complete a weeding task of about 200m 2 per hour,and the system’s success rate of obstacle avoidance reaches 98%.In the actual test,the robot can successfully avoid crop stones,road undulation,and at the same time can dynamically adjust the path according to real-time feedback,and efficiently navigate in complex terrain.The research in this paper can provide new ideas and solutions for weeding technology in the field of intelligent agriculture,and provide technical support for promoting intelligent and refined management of agriculture.
作者
张克盛
王云德
王雷
刘孜文
ZHANG Kesheng;WANG Yunde;WANG Lei;LIU Ziwen(Gansu Polytechnic College of Animal Husbandry and Engineering,Wuwei 733006,China)
出处
《农机使用与维修》
2025年第8期70-73,共4页
Agricultural Machinery Using & Maintenance
关键词
智能除草
计算机视觉
深度学习
路径规划
农业机器人
intelligent weeding
computer vision
deep learning
path planning
agricultural robot
