Application of computer vision and support vector regression for weight prediction of live broiler chicken

A very important ingredient in the recipe for a productive broiler breeder flock is the collection of frequent and accurate body weights. To achieve this goal in this paper image processing and support vector regression (SVR) were used as a non-invasive method. An ellipse fitting algorithm using generalized Hough transform was performed to localize chickens within the pen and the head as well as the tail of chickens was removed using Chan-Vese method. After that from broiler images six features were extracted, namely area, convex area, perimeter, eccentricity, major axis length and minor axis length. According to statistical analysis between weight estimation of SVR and manual measurement of birds up to 42 days, no significant difference was observed (P > 0.05). The RMSE (root mean square error), MAPE (mean absolute percentage error) and the R2 (correlation coefficient) value of SVR algorithm were 67.88, 8.63% and 0.98, respectively. This shows that machine vision along with SVR could promisingly estimate the weight of life broiler chickens.