Application Of Kernel-PCA And Curve Integration In The Cow Body Condition Score

Cow body condition score is an important assessment method to use to measure the energy metabolism of the cow, which is primarily based on body fat deposition and widely used in dairy management. The traditional cow body condition score need expert rating score by touch judge and expert visual assessment combined based on scoring rules and professional experience. This scoring method is not only a higher dependence on professionals, time consuming, and there is some subjectivity differences.This paper presents two automatic body condition scoring method based on image processing techniques. Researching cow body condition score of rules and standards, extracting the main physical characteristics affecting scores of cows. Finally, using digital image processing and pattern recognition technology for automatic body condition score cows.Digital image processing is a methods and techniques to process image by computer to remove noise, enhancement restoration, segmentation, feature extraction, which widely used in military, meteorology, and medicine. Pattern recognition is a process to analysis and processing, extracting which can describe and identify objective thing. Pattern recognition is an important part of artificial and intelligence. Image processing and pattern recognition technology has been successfully applied in areas such as face recognition which is very mature now. There have no standardized cow picture library is available in the application of automatic cow body condition score. The application of image processing and pattern recognition technology in cow body condition score in infancy, scoring accuracy is not satisfactory. For lack of research in this field, this paper focuses on application research of body condition score by using of image processing techniques. Shooting cow images, recognizing score to establishing a standard cow body condition score library.The main contents of this paper are the following:(1) Reading cow body condition score and linear scoring criteria and research, combined with expert rating score cows experience, careful analysis of body condition score and body linear score of assessment methods, the main part, scoring key points. Understanding contribution of the various characteristics of cows to the overall score and to determine scoring site of cows. To visit and study to cow breeding base, collecting cow buttocks behind pictures and record professionals cow body condition score, provide learning material for future trials.(2) Reading inspection papers of Face Recognition, summarize the process of recognition, feature extraction algorithm and feature matching algorithm, learn from the successful experience of face recognition, analysis of the status quo of domestic and foreign auto body condition score of cows, proposed two automatic cow body condition scoring method based on image recognition--the hip curve integral method and Kernel principal component analysis.(3)Test simulation phase, select smooth surface and clear cow picture to pretreatment. Hip curve integration method needs to cut off background of image, then extract the hip characteristic curve and integral to it, with the level of the standard data values obtained by comparison of cow body condition score draw appropriate. Compare integration with standard score values. Kernel-principal component analysis method requires to remove friesian of cow, and lock the root zone of cow. Then using of Kernel-principal component analysis to extract cow body condition statistical characteristics, and then compare with characteristics data of the innovation of this paper can be summarized as the following three points:(1) In the image pre-processing, innovative apply Hough transform to cow tail region extraction innovatively, which can promote research of cow body condition automatic score;(2) Introduced Kernel-principal component analysis into cow body condition automatic score, improved accuracy of cow body condition score;(3) Hip curve integral method which obtain score by compared integration with standard cow body condition score (called herein as the hip curve integral method). Integration can be gotten by to integral to extracting characteristic curve of the hip. This method greatly reduces the computational complexity of scoring, which will helpful to the promotion and application of automatic scoring.