基于卷积算法的自动割草机目标检测应用研究  

Research on the Application of the Target Detection of Automatic Mower Based on Convolution Algorithm

作　　者：叶继阳[1] Ye Jiyang(Jinhua Polytechnic, Jinhua 321007, China)

机构地区：[1]金华职业技术学院,浙江金华321007

出　　处：《农机化研究》2022年第6期227-231,共5页Journal of Agricultural Mechanization Research

基　　金：浙江省科技计划项目(2017C54001)。

摘　　要：为进一步提高我国自动割草机的目标检测准确度与整机作业效率,采用卷积算法理论针对其检测系统展开设计应用研究。在整机结构组成及作业原理基础上,根据卷积算法执行规则流程,建立目标函数与检测核心模型,并针对检测系统进行软件设计与硬件平台搭建,并进行自动割草机目标检测作业试验。结果表明:经卷积算法应用,自动收割机检测系统可实现目标的快速分类,目标识别准确率平均可达96.33%,图像识别清晰度可提高至95.90%,漏割率大幅度降低,作业效率可提升10%以上,可为割草装备的精准深度优化提供一定的改进思路,且对农业图像视觉识别学科的拓展好具有重要的现实意义。In order to further improve the accuracy of target detection and the whole efficiency of the automatic mower in China,the convolution algorithm theory was used to design and apply the detection system.On the basis of the structure and operation principle of the whole machine,according to the execution rule flow of convolution algorithm,the target function and the core model of detection were established,the software design and hardware platform were built for the detection system,and the target detection operation test of automatic mower was carried out.The results showed that the automatic mower detection system could realize the target fast classification through the application of convolution algorithm,the target recognition accuracy could reach 96.33%on average,and the image recognition clarity could be increased to 95.90%.The automatic mowing leakage rate was greatly reduced,and the operation efficiency of the whole machine could be increased by more than 10%compared with the previous.The system application design was feasible,which would provide some improvement ideas for the precise depth optimization of the mowing equipment,and also would be a very important practical significance for the development of agricultural image visual recognition discipline.

关 键 词：自动割草机 目标检测 卷积算法 识别准确率 漏割率 

分 类 号：S817.111[农业科学—畜牧学]