基于图像处理技术的稻田杂草盖度计算研究  

Weed coverage calculation in paddy field based on image processing technique

作　　者：韩正笑 刘涛 陈瑛瑛 孙成明 武威 陈雯 

机构地区：[1]扬州大学江苏省作物遗传生理国家重点实验室培育点/粮食作物现代产业技术协同创新中心,江苏扬州225009

出　　处：《中国农机化学报》2015年第4期193-196,共4页Journal of Chinese Agricultural Mechanization

基　　金：国家级大学生创新创业训练计划项目(201311117036);江苏省高等学校大学生创新创业训练计划重点项目(201311117036Z);江苏高校优势学科建设工程资助项目

摘　　要：针对稻田杂草群落分析中杂草盖度指标采集困难的问题,研究利用图像处理技术对稻田杂草盖度测算的方法。利用超绿特征值和最大类间差法获取田间绿色植被二值图像,通过多次腐蚀并清除较小区域的操作将杂草与栽培作物分离,再通过盖度计算公式求出杂草盖度。对样方图像进行盖度测量实验发现,利用图像处理方法计算的杂草盖度与方格法计算的结果平均偏差仅为0.512%。结果表明,图像处理方法可以实现复杂环境稻田杂草群落分析中杂草盖度的计算。Aiming at the difficulty of collecting weed coverage indicators in the weed community analysis,a method based on image processing technology was designed.Firstly,extra-green character and OSTU algorithm were used to get binary image of green vegetation.Secondly,weeds and crops were separated from each other by repeatedly corroding and removing smaller regions in binary mathematical morphology,and then weed coverage was calculated again by corresponding equation.Graticulation method was used to test the sample image,and the deviation between it and that using image processing method was only 0.512%,which shows the image processing method could realize weed coverage calculation for complex weed community.

关 键 词：稻田 杂草 盖度 图像处理 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]