Automatic Swine Detection and Counting Using Hybrid Filter and Ellipse Fitting Model

Imaging and database technology has made business owners to be motivated to integrate automation to their business operation (i.e. from small businesses to large enterprises). Swine or hog raising, for example, is a very popular enterprise in the Philippines, whose challenges in production monitoring can be addressed through technology integration. Swine production monitoring can become a tedious task as the enterprise goes larger. Specifically, problems like delayed and inconsistent reports are most likely to happen if counting of swine per pen of which building is done manually. In this study, we present smartCount, which aims to ensure efficient swine detection and counting that hastens the swine production monitoring task. We introduce a method that automatically detects and counts swine based on Sobel filter with 8-neighborhood implementation and ellipse fitting model, given the still photos of the group of swine captured in every pen. Furthermore, the system can generate periodic production reports and can identify the specific consumables to be served to the swine according to schedules.