基于色差信息的田间成熟棉花识别  

Study on the recognition of mature cotton based on the chromatic aberration in natural outdoor scenes

作　　者：王勇[1] 沈明霞[1] 姬长英[1] 

机构地区：[1]南京农业大学工学院,江苏南京210031

出　　处：《浙江农业学报》2007年第5期385-388,共4页Acta Agriculturae Zhejiangensis

基　　金：教育部重点科研项目(03091)

摘　　要：准确识别和定位目标在三维空间中的位置,为机械手提供运动参数,是采摘机器人视觉系统要解决的关键问题。本文利用棉花成熟期,棉花表面颜色与背景颜色存在较大差异的特征,在统计分析的基础上,建立了利用色差信号进行棉花识别的视觉模型。试验结果表明该模型可以完成棉花的识别,识别准确率达到85%以上。Accurate recognition and localizing target in the three-dimensional space position that provide the movement parameters for the manipulator are the two key problems of improving the vision system of harvesting robot. According to the characteristic difference between the cotton surface and the background color during cotton maturity, this paper established the visual module that recognize the mature cotton in natural outdoor scenes by using the chromatic aberration information based on the statistical analysis with SPSS 10.0 for Windows. The testing results indicated that the module could recognize the cotton under the complex background. The accurate rate of recognition was more than 85 % .

关 键 词：机器视觉 棉花 图像处理 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]