Image Segmentation of Cows using Thresholding and K-Means Method

Abstract

Cow's weight parameter depends on the characteristics and size of the cow's body. This system aims to segment body parts of cows using thresholding and K-Means method to produce cow body extraction as an early stage in the process of estimating cow's weight. The thresholding method begins by inputting a digital image then performing a sharpened grayscale process with edge detection and dilation processes. As a comparison, segmentation with K-Means method would segment the image into two (2) clusters. The results showed better segmentation of cow's body with local thresholding method than the other two methods.