害虫自动识别与计数技术研究进展与展望  

Research progress and prospect of technologies for automatic identifying and counting of pests

作　　者：陈梅香[1,2] 杨信廷[1,2] 石宝才[3] 李文勇[1,2] 杜晓伟[1,2] 李明[1,2] 孙传恒 

机构地区：[1]国家农业信息化工程技术研究中心,北京100097 [2]农业部农业信息技术重点实验室,北京100097 [3]北京市农林科学院植物保护环境保护研究所,北京100097

出　　处：《环境昆虫学报》2015年第1期176-183,共8页Journal of Environmental Entomology

基　　金：北京市自然科学基金项目(4132027);国家自然科学基金项目(31301238)

摘　　要：害虫种类的识别与发生数量的获取是精准预测的重要基础,对害虫综合治理具有重要的决策意义。传统的害虫识别与计数方法效率低,难以满足植保现代化的要求。本文综述了害虫自动识别与计数技术的研究进展,讨论了声音信号、图像技术、红外传感器在害虫自动识别与计数中应用的优势和不足,对提高害虫自动识别与计数技术准确率的解决措施提出了初步的建议。提出今后研究重点应将上述多种技术进行结合,不同技术之间相互取长补短,建立综合性的监测技术体系,以提高害虫识别与计数的准确率;同时应用三维技术解决害虫二维图像识别算法普适性的问题。Getting pest species and quantities are the important foundations for pest accurate prediction, and have a great significance for integrated pest management. The efficiency of traditional method is lower for identifying and counting pest, which can not meet the actual needs of modern plant protection. This review describes the research progress of technologies for automatic identifying and counting pests, analyzes the advantages and disadvantages of acoustic signal, image technology, infrared sensors for the automatic identifying and counting pests, and proposes the main solution and future research priorities to improve the techniques identifying and counting of pests. The various monitoring techniques should be combined to establish comprehensive monitoring technology to improve the pest identifying and counting accuracy, and three -dimensional （3D） technology should be used to resolve the universality of pest identification algorithm based on two- dimensional （2D） image information.

关 键 词：害虫 自动识别 计数 红外传感器 声音信号 图像技术 

分 类 号：Q968[生物学—昆虫学] S435[农业科学—农业昆虫与害虫防治]