Hog Weight Estimation using Image Processing

Abstract

The demand for food rises proportionally as the population grows. To be able to achieve a sustainable supply of livestock products, efficient farm management is a necessity. In the Philippines, pig production, food processing, and profitability ratios can be detected by identifying the live weight of live pigs in real-time. Traditional pig weight detection often requires direct contact with pigs, which is limited by low efficiency and can result in even death. Detecting the weight of non-contact pigs has become a challenge in pig production for decades. The advancement in technology also brought innovations that could be harnessed in order to achieve better productivity in animal production and agriculture. Image processing is a promising concept of making use of cameras or available devices to automatically and continuously monitor and manage livestock images for production. With this concept, this paper introduces the integrate image processing technique for weight monitoring using thresholding with the Otsu method, the conversion of an image from true-color image to greyscale image, and the combination of erosion and dilation process to fill in the missing image elements in order to have a better conversion of image to weight. The height of the camera is 6ft. There are 20 hog sample data with the actual weight and the camera image converted weight with an average of -0.041% difference.