基于压缩感知的农作物病虫害图像重构  

Image Reconstruction of Agricultural Plant Diseases and Insect Pests Based on Compressive Sensing

作　　者：华晶[1,2,3] 易文龙[1,3] 何火娇[1,3] 

机构地区：[1]江西农业大学软件学院,南昌330045 [2]南昌大学机电工程学院,南昌330031 [3]江西省高等学校农业信息技术重点实验室,南昌330045

出　　处：《农机化研究》2015年第10期72-75,共4页Journal of Agricultural Mechanization Research

基　　金：国家自然科学基金项目(61363041);江西省教育厅青年基金项目(GJJ11090)

摘　　要：针对农业病虫害远程诊断系统中对农作物病虫害图像重构精度要求高的问题,提出基于压缩感知的病虫害图像重构方法。利用压缩感知方法,在传感器数据采集端使用测量矩阵对病虫害图像信号进行观测,测量值通过网络被传输至远程监控端后,再使用正交匹配追踪算法对图像进行重构。为了验证提出的重构方法的有效性,在不同的采样数量下,对病虫害图像的重构进行了仿真。结果表明,该方法能高质量地重构病虫害图像。Regarding to tackle the problem of demanding high accuracy of reconstructive agricultural plant diseases and insect pests image , a method of diseases and insect pests image reconstruction based on compressive sensing is proposed . Based on the compressive sensing method , it uses measurement matrix to measure the diseases and insect pests image sig-nal on the sensor nodes .After the measured value is transmitted to the remote monitoring by network , orthogonal matc-hing pursuit algorithm is used to reconstruct the image .In order to illustrate the effectiveness of the proposed method , the diseases and insect pests images under different sample quantity are simulated .Experimental results show that the pro-posed method can reconstruct the image with high quality .

关 键 词：病虫害图像 重构 压缩感知 正交匹配追踪 

分 类 号：S43[农业科学—农业昆虫与害虫防治] TP919[农业科学—植物保护]