基于机器视觉的茶园害虫智能识别系统研究与实现  

Research and Implementation of Intelligent Recognition System for Tea Garden Pest Based on Machine Vision

作　　者：潘梅 李光辉[1] 周小波[1,2] 阮红丽 卢珍[1] 曾文明[1] 李玉玲[1] 赵九洲[1] 梁君 PAN M ei;LI Guang-hui;ZHOU X iao-bo;RUANHong-li;LU Zhen;ZENG Wen-ming;LI Yu-ling;ZHAO Jiu-zhou;LIANG Jun(Sichuan Research & Design Institute of Agricultural Machinery, Chengdu Sichuan 610066;Key Laboratory of Water Saving Agriculture in Hill Areas in Southern China of Sichuan Provincial)

机构地区：[1]四川省农业机械研究设计院,四川成都610066 [2]南方丘区节水农业研究四川省重点实验室

出　　处：《现代农业科技》2019年第18期229-230,233,共3页Modern Agricultural Science and Technology

基　　金：公益性行业(农业)科研专项经费项目(20150312704);四川省科研院所科技成果转化资金项目(2018YSZH0030);四川省基本科研业务费项目(2018JBKY0005)

摘　　要：针对茶园害虫依靠人工识别方法存在效率低、需人工干预的局限,不利于茶叶植保过程中害虫的自动识别及精准施药的信息化、机械化的问题,提出采用机器视觉技术实现茶园害虫的智能识别。本文对其技术路线、处理流程、研究与实现平台和方法进行研究,采用HSV空间的阈值分割、SIFT特征提取、SVM分类等算法实现茶园害虫智能识别系统,并以茶尺蠖为例验证了该识别系统研究与实现方法的可行性和有效性。结果可为茶叶植保过程中的自动、精准施药提供依据,促进农业机械向着智能化方向发展。Aiming at the limitation of low efficiency and manual intervention by artificial recognition method in the tea garden pest recognition,which is not conducive to the automatic recognition of pests and the informationization and mechanization of precise pesticide application in the process of tea garden protection,the intelligent recognition of pests in the tea garden based on machine vision technology is proposed and used.In this paper,the technical route,processing flow,research and implementation platform and method are studied.The intelligent recognition system for tea garden pests is realized by threshold segmentation in HSV space,feature extraction using SIFT,classification by SVM and other algorithms.The feasibility and effectiveness of the research and implementation method are verified by taking the tea inchworm as an example.This can provide basis for automatic and precise application of pesticides in the process of tea plant protection and promote the development of agricultural machinery towards intelligent direction.

关 键 词：机器视觉 茶园害虫 智能识别系统 精准施药 植保机械化 

分 类 号：S24[农业科学—农业电气化与自动化]