Enhancing Precision Livestock Farming Management with AI-Driven Ear Tag Detection and OCR Recognition

Automated cow identification is crucial for Precision Livestock Farming (PLF) which aims to improve livestock production, reproduction, health, welfare, and impact on the environment positively. Automating identification of individual cows using sensor systems such as Radio-Frequency Identification (RFID) and Quick-Response (QR) code is effective but costly, particularly for medium and large farms. To ensure sustainable cow management, cost-effective, plug-and-play systems are needed. In this context, we introduce an accurate, low-cost method for cow identification using a single RGB camera. Utilizing state-of-the-art real-time object detection with YOLOv8, our approach detects and recognizes cows by their ear tag numbers. The cow IDs images are pre-processed and then read by an Optical Character Recognition (OCR) algorithm. The experimental results demonstrate the effectiveness of our method, achieving an ear tag detection mAP@0.5 of 91.60% and a 92% improvement in number recognition accuracy through the proposed image preprocessing.