Automatic geometric parameter measurement for beef cattle grouping platform

To overcome the limitations of low efficiency and insufficient accuracy in traditional geometric feature measurement of beef cattle, we have developed a novel intelligent grouping platform that integrates deep learning and depth camera technology. The proposed approach employs a RealSense D455 depth camera to capture both RGB and depth data of cattle, utilizes DeepLabCut for precise keypoint detection, and calculates geometric parameters through a dynamic depth replacement optimization method. Experimental results demonstrate that the platform achieves measurement errors of less than 2.5% for key metrics, including body length, body width, body height, hip width, and hip heigh, which are critical for accurate size-based grouping in livestock management. With a processing speed of 6 FPS, our platform effectively meets the requirements for real-time and efficient measurement in large-scale farming operations. This study provides robust technical support for precision livestock grouping, digital management and smart breeding practices in modern intelligent farms, marking a significant advancement in precision livestock measurement technology.