基于光谱的作物病害识别研究进展-定稿导论.doc

专业文献综述 题 目: 基于光谱的作物病害识别研究进展  姓 名: 周红娅 学 院: 信息科学技术学院 专 业: 计算机科学与技术 班 级: 计科122班 学 号: 指导教师: 姜海燕 职称: 教授  2015 年06月28日南京农业大学教务处制 基于光谱的作物病害识别研究进展 作者：周红娅 指导老师：姜海燕 摘要：目前我国的粮食生产每年都因作物病害造成巨大经济损失。传统的作物病害识别方式耗时耗力，已满足不了当前对作物病害识别的需求。本文对传统的作物病害识别方法进行了简单的叙述，综述了基于高光谱、多光谱技术的多种作物病害识别的研究进展，病害有番茄灰霉病、黄瓜白粉病、水稻稻瘟病等，以及介绍了每个病害识别研究中对光谱数据进行处理的各种算法。光谱技术在农作物病害识别方面虽得到广泛应用，但仍存在作物、病害种类不全，以及检测的区域有限等问题，最后对光谱在病害识别发展前景进行了展望。 关键字：作物病害；高光谱；多光谱；病害识别 Research Progress on Crop Diseases Recognition Based on Spectrum ZHOU Hong-ya，JIANG Hai-yan (Nanjing Agricultural University, College of Information Science and Technology, Jiangsu Nanjing 210095) Abstract: At present, Chinas food production has huge economic loss every year because of crop diseases. The traditional methods of crop diseases recognition are time-consuming and power-consumption and cant meet current demands for crop diseases recognition. The traditional methods of crop diseases recognition and the research’s progress of crop diseases recognition based on hyperspectral and multispectral technology are introduced in this paper.These diseases including tomato gray mold ,cucumber powdery mildew and rice blast of rice are researched.Various algorithms for processing spectral data are summarized.Spectral technology has been widely used in crop diseases recognition,but there still exists some problems including incomplete crop varieties,incomplete diseases and limited detection areas.Finally, the prospect of the development of the crop disease recognition based on Spectrum is discussed. Key words: crop diseases；hyperspectral；multispectral；diseases recognition 引言：作物病害是造成粮食生产损失最大的因素，而且农作物病害的种类繁多，有1400种之多，成为作物生产最大的隐患。中国是农业大国，只用了7%的土地就养活了全球20%的人口，所以农作物病害对中国农作物生产的影响不容小觑，如果能做到在病害发生或发生之初就识别检测到作物病害