| There are widespread problems such as high stocking density,short growth cycle and poor welfare in broiler breeding in China,which will seriously affect the mobility and health status of broilers,thus restricting the welfare of broiler and economic benefits.Based on the broiler gait characteristics and artificial rating standards of broilers,this paper developed a broiler leg defect identification and early warning system for the problem of low manual evaluation of broiler gait scores and difficulty in identifying early broiler leg defects.During the industrialized breeding process of broilers,the growth of broiler’s legs cannot be matched with its rapid weight-gaining rate.Coupled with poor environmental conditions and extensive management methods,the incidence of severe leg diseases in the broiler life cycle reaches nearly 20%.The pathogenesis of broiler leg disease is usually slow,and no significant treatment can be given until significant conspicuous symptoms occur,resulting in huge losses.Therefore,it is of practical significance to evaluate the health status of broiler legs in a timely and accurate manner by means of automated method and to alert the possible leg diseases,which will improve the welfare level and economic benefits of broiler breeding.This paper mainly completed the following aspects:(1)Design a broiler walking experimental platform.In order to solve the problem of lack of experiment environment,a broiler walking platform integrating RFID individual identification and broiler video collection function was established to realize automatic cutting and preservation of single broiler through video.(2)Research the individual target detection methods for broilers and broiler dynamic parameter extraction method.In order to realize the recognition and localization of broiler targets under different postures and behaviors,based on the deep convolutional neural network,a broiler individual target detection method was proposed and proposed,which can realize the target recognition for broilers from complex environments.The accuracy of target recognition under natural conditions was 93.39%,which was close to 100%under laboratory conditions.The identified broiler area is used as the ROI,and the moving variables such as the speed,stride length and stride length of the broiler are extracted by image processing methods.(3)Define the abnormal index of walking parameters of broilers.In order to eliminate the individual differences and provide support for evaluating the walking status of broilers,the ideal movement parameters of broiler gait are defined according to the elliptical fitting area of the broiler.The moving parameter anomaly index is defined according to the difference of actual movement parameters and ideal moving parameters of the broiler,which is used as the reference for rating the broiler gait score.(4)Establish an automatic evaluation model for gait score.Based on the moving abnormal index,an automatic broiler-gait evaluation method were proposed by decision tree model.The average accuracy of gait scores is 78%,and the optimal accuracy rate is 88%.Identification and early warning of broiler leg defects are based on broiler gait scores.(5)Develop WEB broiler leg defect information management system.In order to achieve the broiler gait video and evaluation results of the acquisition,transmission,processing,display,storage and query functions.Develop a B/S architecture broiler leg defect information management system.In this paper,the 40-day-old yellow-feathered broiler was taken as the object.The video of standardized state of broiler gait was collected.The broiler target was identified and located in real time through deep learning model and target recognition method.The image processing algorithm was used to separate and extract the broiler,extract the moving parameters such as speed,stride length,difference of stride length and stride frequency of the broiler,finally realize the automatic evaluation of gait score of the broiler based on the decision tree model,which supports for the upgrading of broiler industry and the improvement of animal welfare. |
