Sow Localization in Thermal Images Using Gabor Filters


In this paper, we focus on the problem of detecting pigs. Since there are many different animals manually identifying them can be a difficult task, this algorithm classifies Pigs based on their images so we can monitor them more efficiently. Specifically, we propose a two-step approach for sows recognition. In the first step, we build a model that can distinguish pigs apart from other animals and humans by using Supporting Vector Machine (SVM), which should correctly classify the Pig based on the Histogram of Oriented Gradients (HOG) feature. After creating this model, we used Gabor filters concatenated with the (HOG) feature for more effective utilization in the second step. We demonstrate that our two-step approach can result in a significant improvement. The results show that the proposed system can detect pigs even in low rank images and has an excellent performance efficiency. Furthermore, Pig detection and classification can assist us in determining when pigs are ready to breed by optimizing heat detection in sows and gilts.

Publication Title

Lecture Notes in Networks and Systems