Design And Research Of Dairy Cow Early Limp Recognition System

Dairy cows' lameness are widespread in dairy farms and have become the second largest category of diseases affecting the health of the dairy industry.It has a great influence on milk production and fecundity of dairy cows.And early lameness is not only difficult to detect but also easy to miss the best treatment time,causing serious economic losses to the pasture.So early detection and early treatment is the key to solve the problem.The pressure measurement gait acquisition system(Walkway)of our group can collect the original gait pressure data,but it is very difficult to extract and calculate the multigait parameters and identify the gait because of the lack of subsequent data processing software.Therefore,it is necessary to design a dairy early lameness identification system for dairy cows based on machine classification algorithm.In this paper,based on MATLAB,the whole design of cow early lameness identification system is completed,and the system analysis and detailed design were carried out.The lameness recognition system mainly includes several modules,such as data reading and saving,data processing and analysis,graphics and curve drawing and discriminant analysis,etc.The detailed design mainly includes the GUI interface design and the detailed programming design of each module of the lameness recognition system.The module of discriminant analysis uses the combination of principal component analysis and Fisher discriminant analysis to establish the lameness discriminant model.In this paper,the pressure measuring gait acquisition system was calibrated by using the cow gait simulator,and the early limp identification system was studied based on three sampling frequencies: 50 Hz,100Hz and 200 Hz.The gait acquisition system and the early limp identification system were used to compare and analyze the experiment.Using this early identification system,several groups of subjects were tested.The test results were composed of sample data by means of artificial gait scores.The training discriminant model was trained and tested.The experimental results show that the error of the gait parameters processing efficiency is within 5% for the three different frequencies in the early identification system,and the comprehensive treatment effect is the best under the sampling frequency of 50 Hz.The gait characteristic parameters are processed by the data processing module.Errors are guaranteed between 0.0% and 3.64%.The pressure curve diagram drawn by the curve plotting module is consistent with the trail collection system.The accuracy rate of the discriminant model trained in the discriminant module is 96.7%.The results of the regression analysis of the training model are both Consistent with the training results,the accuracy of the test sample was 100%.The early identification system of the cows has high accuracy in gait data processing.The curves are drawn accurately and can be used to identify cows in the early stages and meet the design requirements of the system.