Classifying Major Rice Weeds in the Philippines Using Image Processing and Machine Learning

Weeds are one of the most pervasive problems in rice cultivation, significantly affecting crop yields and contributing to agricultural inefficiencies. This study introduces a novel mobile-based solution utilizing convolutional neural networks (CNN), particularly the MobileNetV2 architecture, to detect and classify major rice weeds endemic to the Philippines. The system integrates image processing, machine vision, and offline database logging within an Android mobile application. By leveraging real-time RGB image analysis and deploying the model on smartphones, this solution provides farmers with actionable insights into weed management. The proposed system demonstrates significant performance with an average classification accuracy of 85.37% under varied field conditions, suggesting a scalable, accessible, and low-cost digital agriculture tool.