Perching Behavior Of Chickens And Development Of An Automated Weighing System For Weight Monitoring Of Group Housed Chickens

Real-time acquisition of body weight(BW)information is essential for precision livestock management.The BW of chickens in commercial farming can be used to know the chicken growth condition and determine the marketing time.However,weighing by hands is inefficient and easy to lead stress to chickens.Due to the perching habit of chickens,the BW information can be automatically sampled when perching behavior happens.However,acquisition of accurate flock BW information is based on the equal perching probability for all chickens,thus the sampled weight data is not skewed towards a lighter or heavier result.The relationship between BW and perching behavior remains unclear,so the feasibility of accurate flock BW monitoring by taking advantage of the perching behavior has not been validated.The opinions on how keel bone(KB)damage or social status influence BW and perching behavior have not reached an agreement,so whether KB damage or social status will influence the accuracy of BW monitoring result is not clear.In this study,an RFID-based automated individual perching behavior monitoring system(APMS)was developed.The APMS was used to investigate relationship between BW and perching behavior as well as the impact of KB damage and social status on BW sampling.Finally,an automated perch-style weighing system was developed.Through the industry-university cooperation,the system was further improved and tested in commercial farms.The results may strengthen the cognition of the perching habit of chickens and provide technical support for the real-time BW monitoring of chicken flocks.The main research contents and conclusions are as follows:(1)RFID-based automated individual perching behavior monitoring systemTo accurately characterize the perching behavior of chicken individuals,an RFID-based automated individual perching behavior monitoring system(APMS)was developed.When a chicken jumped on the perch(30 cm high),the ID information of the chicken in the RFID tags was collected by the RFID reader.At the same time,load cells at both ends of the perch sampled the weight on the perch.The weight information was used to correct the errors caused by RFID detection miss.The perching behavior of the experimental chickens(5 birds/group,35 weeks old)in three groups was monitored and then compared with the result of manual observation,the error rates of APMS on the number of perch visit(PV)and the perching duration(PD)were 0.50%and 0.26%,respectively.It indicated that APMS can accurately characterize the perching behavior of individual chickens,thus can be used in the following experiments to explore the impact of different factors on BW sampling of chicken flocks.(2)The relationship between BW and perching behaviorThe relationship between BW and perching behavior was investigated,thus the feasibility of accurate flock BW monitoring by utilizing perching behavior was validated.The perching behavior of 15 chickens(“De Kalb White”,laying hen,35weeks old,1.54-1.89 kg)was monitored for a week by APMS.The individual BWs were weighed manually.No significant correlation was found between BW and perching behavior(p>0.05).The results indicated that for healthy and active chickens like layers and slow-growing broilers,when sufficient resources and low perches were provided,the perching behavior will not be influenced by BW.Thus,it is feasible to take advantage of perching behavior to realize accurate flock BW monitoring.(3)The impact of KB damage on the automated BW sampling of chicken flocksThe relationship between KB damage,BW and perching behavior of individual chickens was investigated.The perching behavior of 36 chickens with different KB damage severity(“De Kalb White”,laying hen,about 56 weeks old,1.31-2.02 kg)was monitored for a week by APMS.The results showed that KB deformation had a significant positive correlation with BW(p<0.01),but had no significant correlation with PV or PD(p>0.05);The fracture numbers in the middle section had a significant positive correlation with the daytime perching bout duration(PD/PV,p<0.01),but the number of KB fractures and BW had no significant correlation(p>0.05).The results indicated that the KB deformation did not influence the BW and perching behavior at the same time,so as the KB fractures.Thus,the KB damage will not significantly influence the automated BW sampling result of chicken flocks.(4)The impact of social status on the automated BW sampling of chicken flocksThe influence of social status on BW and perching behavior of individual chickens was investigated.Ninety chickens(“De Kalb White”,laying hen,about 40weeks old,1.31-1.98 kg)were assigned into 6 experimental pens with sufficient perch space(17 cm/chicken)and 6 pens with insufficient perch space(8.5 cm/chicken).The perching behavior of individual chickens was monitored for a week by APMS.The individual social status was determined by the difference between each chicken's number of aggressions given and aggressions received,and then the chickens were divided into 3 levels called the strong,the medium,and the weak(rankings of the top20%,20-80%and the last 20%).The results showed that with sufficient perch space,the weak had significantly lighter BW than the flocks average BW(p<0.05),at the same time had significantly higher daytime PD than the strong and the medium(p<0.01).It indicated that the weak was lighter and perched more,which would cause BW sampling data skewed to a lighter result.However,it is rare for commercial farm to equip sufficient perches on floor.With insufficient perch space,which was more closed to the commercial condition,the weak had significantly lighter BW than the strong(p<0.05),but no significant difference in perching behavior among all the 3levels(p>0.05),thus it indicated that the social status will not significantly affect the automated BW sampling result of chicken flocks in commercial farm.(5)An automated perch-style weighing systemAn automated perch-style weighing system was developed to monitor the average BW and uniformity of chicken flocks.When a chicken jumped on the perch,the load cells under both ends of the perch recorded the stable BW on the perch,and the camera above took a picture and the number of chickens on the perch was detected.The BW and the chicken number were recorded to be one BW sample.An algorithm was developed to delete the redundancy and outliers in the BW samples.Nine BW data from two broiler groups(“Xueshan Grass Chicken”,hens,9-13 weeks old,100 birds/flock,11.7 chickens/m~2)were obtained by the weighing system.Optimized parameters of the data processing algorithm were derived from the first 5BW data and applied to the rest 4 BW data.Compared with the true weight information,the error rates of the automated weighing results for average BW and flock uniformity were–1.00-2.17%and–0.99-1.57%,respectively,and better than manual sampling weighing method.This automated perch-style weighing system had been further improved through industry-university cooperation and tested in 62Xueshan Grass Chicken flocks in commercial farms.The error rate of flock average BW was±1.68%,which indicated the system's performance was accurate and stable.