基于FISHER模型的棉种分色研究  

The Color Srting of Cotton Seed Based on FISHER Model

作　　者：洪英[1] 邓继忠[2] 孙莎[2] 杨军[3] 刘跃[3] 

机构地区：[1]天津工业大学机电学院,天津300160 [2]华南农业大学工程学院,广州510642 [3]新疆农科院农机化所,乌鲁木齐830091

出　　处：《农机化研究》2008年第12期113-115,118,共4页Journal of Agricultural Mechanization Research

基　　金：新疆少数民族科技骨干人才特殊培养科研专项资金项目(200623110)

摘　　要：利用棉花种子表皮光学特性,对颜色不正常、表面有缺陷或内部变质的疵品及杂物进行分选是一种有效的方法。研究棉种分色模型是研制自动化棉种分色机械的基础。为此,通过提取彩色棉种图片的RGB分量,利用Fisher准则对RGB分量特征值进行提炼,建立了棉种的分选模型。试验表明,该色选模型可以对棉种进行有效分选,其分选精度达到97.2%。The cotton seed capsule usually appears different colors because of the maturity. It appears deep brown or black when it is maturate and red when it is not,so a way to assort cotton seed is designed by the method of image processing. The RGB eigenvalue was extracted from the chromoimage of cotton seed, and then a sorting model was built by the Fisher arithmetic. Test results show that the Fisher model could realize effective identification of the cotton seeds and the correctness could reach over 97%.

关 键 词：棉种 分色 机器视觉 Fisher模型 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] S126[自动化与计算机技术—计算机科学与技术]