基于支持向量机的鹅肉肉色客观评定研究.pdfVIP

基于支持向量机的鹅肉肉色客观评定研究 1 2 *，1 1 1 1 涂冬成 ，薛龙 ，刘木华 ，赵进辉 ，沈杰 ，吁芳 （1江西农业大学 工学院，江西 南昌 330045） （2华东交通大学机电工程学院，江西 南昌 330013） 摘要：为了对鹅肉肉色进行更加快速、客观的评定，搭建了可见光图像采集系统，获取了 98 个新鲜鹅肉 RGB 图像。首先应 用过阈值分割方法去除背景，获取颜色的有效判别部位，鹅胸脯肌肉区域的图像，然后通过对该图像进行颜色空间转换提取 肉色参数 （L*、a*、b*），最后以归一化后的肉色参数作为特征向量，应用支持向量机建立了肉色评定模型。结果表明，训 练集和测试集的准确率分别为 89.0625%和79.4118%，为鹅肉肉色快速、客观的在线评定打下了良好的基础。 关键词：鹅肉; 肉色;过阈值分割;支持向量机;客观评定 Study on objective evaluation of goose color based on Support Vector Machine 1 2 * 1 1 1 TU Dong-cheng ,Xue Long ,LIU Mu-hua ,ZHAO Jin-hui ,Shen Jie ,Yu Fang (1College of Engineering, Jiangxi Agricultural University, Nanchang, JiangXi, 330045, China) (2 School of Mechanical and Electronical Engineering, East China JiaoTong University, Nanchang, JiangXi 330013,China) Abstract:In order to realize a more objective and fast eveluation of goose color, a visible image acquisition system was built to capture 98 RGB images of goose samples. Firstly, Over-threshold segmentation method was applied to eliminate background for extracting effective discrimination area of color, then the fleshcolor features (L*、a*、b*) were extracted through color space conversion, finally the fleshcolor features which had been normalized were used as feature vector to build goose color evaluation model based on Support vector machine (SVM).The results showed that the accuracy rate of traning set and test set are 89.0625% and 79.4118%, respectively,