Research on individual cattle identification algorithm based on QR code

Accurate individual identification of cattle is a prerequisite for intensive, large-scale farming. To address the shortcomings of traditional ear tag identification—namely, easy tag detachment or damage—as well as the high hardware and data requirements of deep learning methods and the frequent need for model updates, this study proposes a QR code–based cattle identification solution. First, we integrate a Coordinate Attention (CA) module into the backbone of the YOLO (You Only Look Once) v8 model to enhance detection accuracy and robustness in complex scenarios. Results show that the mAP reaches 97.6%, maintaining strong detection speed and a manageable parameter count while improving precision. Next, an improved SRGAN super-resolution network is used for image upscaling, leading to improved SSIM and PSNR values, with a Zbar decoding success rate of 92.1%. These findings underscore the approach’s substantial potential for real-world farm management applications.