田间叶螨图像二维LWT小波提升分离及识别  

Isolation and Recognition of Field Spider Mite Image Based on Twodimensional LWT Wavelet

作　　者：张华[1] 刘国成[1] 

机构地区：[1]广州铁路职业技术学院信息工程系,广州510430

出　　处：《科技通报》2014年第8期209-211,共3页Bulletin of Science and Technology

基　　金：广东省自然科学基金培育项目(GTXYP1310)

摘　　要：为了对田间叶螨进行有效采集识别,对传统的基于图像的田间叶螨采集识别方法进行改进,提出一种采用二维LWT小波提升方案的叶螨图像准确分离和识别方法。设计一种基于物联网技术的叶螨图像采集系统,对叶螨病斑区域的超红特征进行灰度化提取,采用二维LWT小波系数对提取的灰度化图像进行提升分离,实现了类病斑区域与非类病斑区域的二值化分离,最后采用小波函数面积阈值重构方案对叶螨进行图像重构,提供给物联网的决策层实现对病虫害的分析决策,实现对叶螨病斑的准确识别。仿真实验表明,采用该算法进行田间叶螨图像识别,图像重构效果较好,对害螨的正确识别率达到96.7%,能有效应用到对田间螨害的实时监测和防治工作中。In order to effectively collect and recognize the leaf mites, the field of the traditional mite collection and identifi-cation method based on image is improved, a lifting scheme mite accurate image separation and recognition methods using dimensional LWT wavelet. A leaf mite image acquisition system is designed based on Internet of things technology, and the gray scale extraction of super red mite lesion characteristics is obtained. It provide decision-making layer to the realization of the Internet of things analysis decision on plant diseases and insect pests, and accurately identify the mite lesion. Simula-tion results show that, by using the algorithm of image recognition field mites, it has good image reconstruction result, and it can correctly identify the mite rate reached 96.7%, it can be applied to real-time monitoring and control of pest mite field.

关 键 词：叶螨 LWT小波 图像识别 物联网 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]