HSV空间下改进的直觉模糊C均值聚类猪肉图像检测  

Improved Intuitionistic Fuzzy C-means Clustering Pork Image Detection in HSV Space

作　　者：王迪 董素芬[1] 程芳[1] 赵艳[1] 李今 WANG Di;DONG Su-fen;CHENG Fang;ZHAO Yan;LI Jin(College of information science and technology,Hebei Agricultural University,Baoding,Hebei 071000,China)

机构地区：[1]河北农业大学信息科学与技术学院,河北保定071000

出　　处：《计量学报》2021年第8期986-992,共7页Acta Metrologica Sinica

基　　金：河北省自然科学基金(60903098)。

摘　　要：利用计算机视觉技术对畜肉分级的方法种类繁多,但由于光照因素使前期图像预处理分割目标和背景的工作难度增大。针对传统的最大类间方差法分割图像效果不佳、噪声适应能力不强的问题,以及核磁共振、高光谱成像等无损检测方法大多存在检测仪器体积大、不便于携带、成本高等问题,提出利用色调、饱和度、明度(Hue,Saturation,Value,HSV)色彩空间结合聚类方法对图像像素点进行聚类分割。在对取自自然光照环境中的猪肉图像进行分割时,所提方法相对于传统聚类分割方法分割正确率平均提高1.46%;在对人工加入了0.1椒盐噪声和0.2椒盐噪声的图像进行分割时,该方法相对于传统方法表现出了更好的抗噪声能力,传统分割方法平均错误率分别升高了16.15%和38.28%,该方法平均错误率仅升高了1.57%和1.49%。该方法具有良好的分割准确率和噪声鲁棒性,提高了目标区域的检测精度,减少了图像预处理阶段的信息丢失,提高了畜肉分级方法的质量。There are many methods to classify meat by using computer vision technology.For the problems of poor segmentation effect and poor noise adaptability of traditional maximum variance(OTSU)method and the problems of large size,unportability and high cost of detection instruments in most nondestructive testing methods such as nuclear magnetic resonance(NMR)and hyperspectral imaging,this paper proposes to use HSV color space image combined with clustering method to cluster image pixel points.HSV color space combined with clustering method is proposed to cluster and segment image pixels.When the pork image is segmented from the natural lighting environment,the proposed method is improved by an average of 1.46%compared with the traditional clustering method.When the image with 0.1 and 0.2 salt and pepper noise is segmented,the proposed method has better anti-noise ability than the traditional method.The average error rate of the traditional segmentation method has increased by 16.15%and 38.28%respectively.The average error rate of the proposed method has only increased by 1.57%and 1.49%.The accuracy of image segmentation and noise robustness improve the detection accuracy of the target region,reduce the information loss in the image preprocessing stage,and improve the quality of the meat classification method.

关 键 词：计量学 畜肉分级 图像分割 C均值聚类 直觉模糊熵 HSV 猪肉检测 

分 类 号：TB96[机械工程—光学工程]