<![CDATA[Milk is a livestock product consumed by individuals of all ages. Therefore, it is essential to increase milk production in Indonesia to meet domestic demand. The growth of dairy cattle populations and milk production has not been able to keep up with rising consumption, resulting in a reliance on imports for most dairy products and their derivatives, with imports steadily increasing over the years. Therefore, alternative solutions are needed to enhance the milk production. One approach is to develop a milk production estimation model to determine the optimal number of dairy cattle to be cultivated by farmers and livestock companies to meet domestic demand. The objective of this study was to create a dairy milk production estimation model through image analysis using the Random Forest, XGBoost, and LightGBM algorithms. The milk production estimation model used in this study used CLAHE for contrast enhancement and VGG-16 for feature extraction. The results showed that XGBoost provided the best performance, explaining 74% of the data variation in the Y variable with a relatively small estimation error of 0.92. After parameter tuning using Grid Search, an improvement was observed, where XGBoost explained 86% of the data variation in the Y variable, and the estimation error decreased to 0.72. Image processing and machine learning technologies are part of precision agriculture that aims to improve the efficiency, productivity, and sustainability of livestock operations.]]>
Copyrights © 2025
