Automatic detection of changes in pig group lying behaviour using image analysis

Nasirahmadi, A; Edwards, S; Richter, U; Sturm, B

doi:10.13031/aim.20152188972

Automatic detection of changes in pig group lying behaviour using image analysis

Lookup NU author(s): Abozar Nasirahmadi, Emerita Professor Sandra Edwards, Dr Barbara Sturm

Downloads

Full text for this publication is not currently held within this repository. Alternative links are provided below where available.

Abstract

Environmental factors provide important information for the better management ofpig farms and they have significant effects on pigs' production efficiency, health and welfare. Due to the physiological and morphological limitations on thermoregulation of pigs, they change their lying behavior to adapt to nigh and low temperatures. In hot conditions they avoid physical contact with others in the pen during resting time and vice versa. Visual monitoring of pig behaviors, usually practiced in small scale farms, is unreliable, expensive and time consuming in large scale farms. The development of image analysis systems could be a reliable and non-intrusive technique for automatic assessment of pig group behavior. The aim of this study was to develop an algorithm for identification of pig group lying behavior under commercial farm conditions using an optical method. Pigs were monitored by a top view CCD camera and animals were extracted from their background using image processing algorithms. The x-y coordinates of each binary image were used for ellipse fitting algorithms to localize each pig. As a result, ellipse parameters such as "Major axis length ", "Minor axis length" and "Centroid" could be calculated for all fitted ellipses. In order to determine the group lying behaviour, the Delaunay triangulation algorithm was applied. By means of the region properties and perimeter of each Delaunay triangule it was possible to automatically find the changes in lying behaviour of grouped pigs and the distance between pigs with high accuracy.