Research On Dorper Sheep Weight Estimation Model Based On Image Processing

In recent years, social economy in our country has developed rapidly, people’s living standard has improved greatly. In this trend, people’s demand for meat is also rising rapidly, it is imminent to realize modernization of large-scale livestock breeding. In the past years, the original manual weighing technology have been used in large-scale livestock breeding, time-consuming and also affects the growth of the sheep. In order to solve the weighing problems in the traditional breeding, the research on Dorper sheep weight estimation model based on image processing has carried on in this thesis.In this thesis, the Dorper sheep was as the main object of the study. Combined with the sheep’s own characteristics and the current status of research, the analysis of the Dorper sheep weight estimation function completed. In this paper, the images of sheep processed by image processing technology, regression analysis method, etc., to obtain sheep body size data, and then analysis of the relationship between body size and weight, to determine the weight model. To complete the research on Dorper sheep weight estimation model based on image processing, what have done in this thesis are as follows:Firstly, the sheep image was processed to eliminate the influence of the scene illumination and noise as far as possible, enhance the contrast between the sheep and the background, then the outline of the sheep obtained after image segmentation. In the image processing of the sheep, different methods have been used, and these methods are compared to select the best solution.Secondly, based on the rules of sheep body measurement in reality, several important sheep body parameters, such as body height, body length and chest circumference, etc., are detected by image processing techniquesThirdly, sample data collected in the field are processed, the regression model between body size and weight are established based on the results. Finally, the estimated body size parameters are substituted into the regression equation, then the study of sheep weight estimation model are completed.Verification results show that the estimate effect of sheep’s weight is more ideal. This method is feasible, and can accurately measure the sheep’s body size and weight. This method is effective to reduce the human and material use, reducing the stimulation of sheep, it is beneficial for the growth of sheep continuous monitoring and research, providing a theoretical basis for improving production management and regulation of feed ratio.