Research On Pig Health Status Evaluation Based On Sound And Visual Fusion

With the large-scale development of China’s pig breeding industry,the traditional operation method of manually observing the health status of pigs has problems such as high labor intensity,low efficiency,and susceptibility to infectious diseases,which can no longer meet the production requirements.Research and development of automated and intelligent monitoring technology is imminent.Therefore,this dissertation takes evaluation of the health status of pigs as the research object,analyzes the manifestations of pig health status,and builds an evaluation index system for pig health status.Based on the analysis of pig behavior and sound characteristics,a pig behavior and call pattern recognition method was developed,a method for evaluating the health status of pigs based on the fusion of sound and vision was established,and the validity of the evaluation method of pig health status in this dissertation has been verified.The research results and main conclusions are summarized as follows.(1)In view of the single evaluation index in the existing pig health state evaluation system,the expression form of pig health state was analyzed,and the behavior and sound of pigs were determined as the evaluation indicators of pig health state,and the pig health state was constructed.The evaluation index system has obtained the overall ranking of the pig health status evaluation index system level.(2)In view of the low efficiency and high labor intensity of traditional manual observation of pig behavior,an algorithm for visual feature recognition of pig behavior was constructed,and the algorithm for visual feature recognition of pig behavior was verified.The recognition accuracy of four common behaviors of pigs,lying on their side,lying on their side and sitting in a dog,all reached more than 94%,which provided technical support for the sound-visual fusion evaluation of pigs’ health status.(3)Aiming at the problem of pig sound recognition,the characteristics of pig sound spectrogram were analyzed,and an algorithm for pig sound spectrogram feature recognition was constructed,and the algorithm was verified.The results showed that the use of convolutional neural network to mine sound spectrogram The deep features can realize the recognition of the pig’s voice.Finally,the Mobile Net V3 deep neural network is selected as the pig’s voice recognition spectrogram feature recognition algorithm model to provide basic data for the sound-visual fusion evaluation of the pig’s health status.(4)Aiming at the quantitative evaluation of the health status of pigs,a sound-visual fusion evaluation method for pig health status was established,and the method was empirically tested.The results show that the pig health status evaluation method based on sound-visual fusion in this paper can Realizing the quantitative evaluation of the health status of pigs can provide technical support for the automatic and intelligent evaluation of the health status of pigs.