| The natural mating cage culture mode is a new type of breeding mode for laying hens,the performance of individual behavior characteristics of laying hens under this cage culture can directly reflect the health status of breeding flocks,realize the automatic identification of individual behavior of breeding hens,and reduce the systematic error of observation and sampling.The automatic identification of individual behavior of breeding hens is the basis of precise breeding management.In this paper,we take the natural mating egg breeder as the test object and designing a real-time monitoring system based on nine axis acceleration sensor and bluetooth wireless transmission,which can automatically monitor and quantify the individual behavior of the chickens in this natural mating cage.Using the wireless acceleration sensor node under the neck of the chicken to collect the acceleration data of the breeder through bluetooth wireless transmission,the system can quickly and continuously collect the behavior information of the breeder in this cross cage,accurately identify the behaviors of the breeder,such as eating,drinking,fighting,mating and flapping,which is helpful to obtain the individual behavior of the breeder without damage and quickly,to provide scientific basis for the exploration of the optimal design and efficient management of the mode of this natural mating cage.The main contents include:(1)An individual behavior collection system for breeders in this natural mating cage is constructed.The system includes an acceleration sensor,a bluetooth adapter,a test breeder,a smart ant camera and PC software.The acceleration sensor transmits the collected data to the host computer software through bluetooth,which guarantees the synchronization between the video and data acquired by the Smart Ant Camera and achieves automatic perception of various behaviors of breeders.(2)Using the method of comparative test,the acceleration sensor is tied to the neck and leg of the breeder to collect data,and the optimal deployment position of the acceleration sensor on the breeder is determined.The results show that the acceleration curve of the legs is not distincted,and the acceleration sensor under the neck can effectively identify five typical behaviors of breeders.(3)A method of detecting individual behavior detection of breeder chicken is proposed based on acceleration sensor,by comparing the five characteristic values of acceleration curve volatility,variance,mean value,extreme value difference and peak value,Wavelet sym denoising is used to preprocessing the original data.K-means mean clustering algorithm is used to identify behavioral features to obtain stable cluster centers and breeder chicken Classification of five typical behaviors,while using video to verify 5 behaviors of breeders.The research results show that the K-means clustering algorithm can accurately recognize the feeding,drinking,fighting,mating,and fluttering behaviors of this breeder breeder.,the average recognition accuracy of feeding behavior is 94.31%,the average recognition accuracy of drinking behavior is 92.53%,the average recognition accuracy of fighting behavior is 84.03%,the average identification accuracy of mating behavior is 72%,and the average identification accuracy of flapping behavior is 92.31%.(4)To carry out the experimental research on the automatic recognition and analysis of the individual behavior of this natural mating cage breeder,The average number of five behaviors of three cocks in this natural mating cage is statistically analyzed,the research results show that the average number of occurrences of feeding behavior of No.1 breeder is seventy-six times,average number of occurrences of drinking behavior is twenty-eight times,average number of occurrences of fighting behavior is twenty-eight times,average number of occurrences of mating behavior is forty-three times,average number of occurrences of fluttering behavior is forty-nine times,a direct observation of the histogram of the seven day average number of behaviors of three breeders,the average behavior of No.1 breeder is the most in fifty-one breeders;the average number of occurrences of feeding behavior of No.2 breeder is seventy-six times,average number of occurrences of drinking behavior is twenty-seven times,average number of occurrences of fighting behavior is thirty times,average number of occurrences of mating behavior is eight times,average number of occurrences of fluttering behavior is twelve times,the average behavior of No.2 breeder is the secondary in three breeders;the average number of occurrences of feeding behavior of No.3 breeder is forty-three times,average number of occurrences of drinking behavior is sixteen times,average number of occurrences of fighting behavior is forty-eight times,average number of occurrences of mating behavior is four times,average number of occurrences of fluttering behavior is two timesthe average behavior of No.3 breeder is the least in three breeders,This system can provide a basis for judging the relationship between typical daily behavior and disease of breeder individuals. |
PDF Full Text Request