论文部分内容阅读
本文以3大类9个稻谷品种为研究对象,运用形态学和多元统计方法,对原始图像进行预处理后,提出了面积A、周长P、长轴L、短轴W、最大半径RMA、最小半径RMI、半径比CR、长宽比CLW和圆形度γ9个稻谷种子形状特征参数的检测算法;利用Bayes判别法建立了籼稻、糯稻、粳稻3大类的类别模型,模型判别效果显著;建立了IRBB10、珍珠矮、培矮64、夷方糯、毫糯量、高杆大糯、中花11、巴利拉、秋光9个稻谷种子品种的品种模型,模型判别效果显著。
In this paper, three rice varieties (9 rice cultivars) were selected as research objects. Morphology and multivariate statistical methods were used to preprocess the original images. The results showed that area A, perimeter P, long axis L, short axis W, maximum radius RMA, Minimum radius RMI, radius ratio CR, aspect ratio CLW and degree of roundness γ9, and established Bayes discriminant method to classify the three types of indica rice, glutinous rice and japonica rice. A variety model of 9 rice seed varieties of IRBB10, pearl dwarf, Pei’ai 64, Yi Fangnuo, Muhuannuo, Dabaoguang, Zhonghua 11, Barila and Qiuguang was established, and the discriminant effect was significant.