Paraconsistent logic used for estimating the gait score of broiler chickens

Abstract

Visually estimating the gait score in a flock is not a precise task. The evaluation is done based on scores from 0 (sound bird) to 5 (lame bird). The extremes are easily identifiable; however, the intermediate scores are not evident. This study aimed to develop an algorithm and software to estimate broiler gait score. Video images were recorded from broilers walking on a special platform. The walkway was covered with bedding substrate. A blue panel provided a background to contrast with the birds. Selected broilers from the different gait scores were placed to walk on the runway, and a video was recorded. An algorithm was developed to analyse the video streams. The centroid of the chicken body was detected, through which the broiler's speed, and acceleration from gait score were calculated. The velocity and acceleration data were analysed using paraconsistent logic, and an algorithm was developed. The software was able to predict the intermediate values of broiler gait score with a low degree of uncertainty, given the broiler velocity and acceleration. For the estimation of GS 1, we obtained 50% accuracy. For GS 2 the estimation was 70% precise, and for GS 3, the results were 100% accurate. During the auditing of the flock welfare process, the intermediate results of broiler gait score are visually difficult to identify. Using the developed software it might be possible to detect lameness in broilers under commercial rearing since the velocity of displacement can be easily measured and used as input data by the growers.