Data Acquisition Of Linear Assessment Cow’s Body Indicators Based On Image Processing

In recent years, with the development of computer vision detection technology, its application in livestock production has gradually increased. The methods of evaluating dairy cow type linear based on image processing become an important part of research about optimization of breeding cows at home and abroad. Evaluation of the Dairy Cow is now mostly carried out manually, and testing equipment and instruments is relatively backward, the use of semi-circular meter, tape measure, calipers and other measuring tools to measure. It has many indicators in evaluation, need to deal with large amounts of data, and there are human factors affect the measurement results.This study used image processing and computer vision technology to complete the measurement of traits of dairy cow. This method avoids contact with the target object, as compared with manual measurement quick and easy, high efficiency, high precision, and avoid subjective factors, while for cows to selection, breeding and breed improvement play a guiding role.The image detection technology is used in dairy cow type linear appraisal. The method of index parameter data acquisition is according to the body characteristics of the various parts of dairy cows to measure. It is more comprehensive, objective, quantitative, and avoids the influence of subjective factors. It has of great significance for dairy production efficiency and economic efficiency.In this paper, the method based on computer stereo vision was used to reconstruct three-dimensional model of cows, then body traits of cows were measured. First, camera calibration was achieved by three-dimensional target, then the SIFT was used for image feature extraction and matching, finally,3D coordinates of feature points were calculated by the projection matrix, to achieve three-dimensional reconstruction of local feature points. For the scope of the camera angle is limited in binocular vision, it was presented that through setting mark points in the adjacent public areas and calculating transformation matrix between different coordinate system according to mark points the local feature points could be converted into a unified coordinate system, then the three-dimensional mosaic of the local area under different viewpoints was achieved. Experiments show that using this method the reconstruction model of dairy cow is relatively good, measurement accuracy and measurement efficiency meet the assessment requirements, and it can be used to replace manual measurement.