42 / 2025-03-31 19:32:11

Yolo-Mobile: A temperature detection model for IoR on body surface of pigs

Infrared thermography,Body surface temperature extraction,Light weight modeling,Machine learning,Critical site recognition

6. 动物健康与疾病控制 > （1）新型动物疾病

Body temperature is a critical indicator of animal health. Therefore, rapid and accurate measurement of pigs' body temperature is essential for effective health assessment. Traditional methods involve rectal thermometer insertion, which is time-consuming and induces stress responses in the animals. Current literature largely emphasizes the accuracy of critical region identification, which increases model size, making it difficult to integrate into edge devices. To address this limitation, this study proposes a Yolo-Mobile model based on infrared thermography, which optimizes the selection of key temperature measurement regions on pigs in infrared images while ensuring the model is lightweight. The results demonstrate that the proposed model achieves an average detection accuracy of 97.9% with a model size of 22.7 MB. Compared to YOLOv9, YOLOv9-c, and Faster R-CNN, the Yolo-Mobile model improves accuracy, recall, and mAP by 3.16%, 4.22%, and 8.43%, respectively. Furthermore, the maximum mean temperature extracted automatically was nearly identical to that extracted manually, with an R² value of 0.989. The method proposed in this study provides a valuable reference for rapid and precise extraction of pig body surface temperatures and health monitoring.