摘要
牛乳体细胞是牛乳质量评价和乳腺炎诊断的一项重要指标。利用图像处理技术对牛乳体细胞进行精准快速的识别,能够为诊断乳腺炎提供更加有效的途径。提出了基于多特征融合与随机森林的牛乳体细胞识别算法,首先从细胞核中提取8种不同的形态特征,然后与细胞的9个颜色特征、16个纹理特征进行特征融合,最后将结果输入到随机森林(Random Forest)分类器中进行特征匹配,识别率达到了95. 75%。
Milk somatic cell is an important indicator of milk quality assessment and mastitis diagnosis. Using image processing technology to accurately and quickly identify could provides effective method for judging mastitis. An algorithm for recognition of milk body cells based on multi-feature fusion and random forest was proposed. First,we extracted 8 morphological features from the nucleus,and fused with the cell’s 9 color features and 16 texture features,then put them into Random Forest( RF) for feature matching.The recognition rate reached 95. 75%.
作者
章潇俪
薛河儒
郜晓晶
周艳青
ZHANG Xiaoli;XUE Hern;GAO Xiaojing;ZHOU Yanqing(College of Computer and Information Engineering f Inner Mongolia Agricultural University ,Hohhot 01001S, China)
出处
《内蒙古农业大学学报（自然科学版）》
CAS
北大核心
2018年第6期87-92,共6页
Journal of Inner Mongolia Agricultural University(Natural Science Edition)
基金
国家自然科学基金项目(61461041)
关键词
牛乳体细胞
图像分割
形态特征
特征融合
随机森林
Milk somatic cells
image segmentation
morphological property
feature fusion
random forest
