粮虫图像识别检测技术研究现状与展望

为促进图像识别检测技术在我国粮情测控领域的应用研究,从预处理、特征提取、识别分类和种群密度估计4个方面,概述和分析了计算机图像识别检测技术在粮虫自动识别中的发展现状,提出今后应从图像自动采集装置的研制、粮虫图像的有效特征获取、高适应分类器的设计、粮虫种群密度的估计、多种检测技术的融合等方面开展深入研究,为科学规范粮虫防治工作提供更好的决策支撑。

Current Research Situations and Prospects of Stored-grain Pests Based on Image Recognition Inspection Technology

In order to promote the application and research of image recognition inspection technology in grain monitoring of China ,current de-velopment situations were firstly summarized for computer image recognition and detection technology applied in automatic identification of stored-grain pests from pre-treatment,feature extraction,classification,and population density estimation.Besides,recommendations were put forward for in-depth study,including developing automatic image acquisition device,obtaining effective characteristics of images of stored-grain pests,desig-ning highly adaptive classifiers,estimating population density of stored-grain pests,and integrating multiple detection techniques,and etc.,so as to provide better support for decision of scientific and standard stored-grain pest control.