IoT-based smart weighing device for swine condition monitoring

The IoT-Based Smart Weighing Device was developed to modernize traditional pig weighing practices by providing a non-invasive, real-time, and accurate method of estimating swine weight. This study aimed to design and implement an integrated system that leverages machine vision, deep learning (YOLOv8), and Internet of Things (IoT) technologies. Using a Raspberry Pi 4 and Arducam camera, the device captures overhead images of pigs and extracts body dimensions — specifically girth and length—through object detection. These features are input into a multi-output regression model to predict weight. The system was developed using the Scrum methodology and includes modules for image processing, object detection, feeding recommendations, and a web-based interface for real-time monitoring. Validation involved unit, integration, and system testing. Statistical analysis revealed strong alignment between actual and predicted values (r = 0.994), a low mean squared error (MSE = 3.703), and no significant difference in T-tests (p = 0.731), affirming model reliability. The software’s quality was evaluated using ISO 25010 standards, while user acceptance was assessed through the Technology Acceptance Model (TAM), which showed high perceived usefulness and willingness to adopt. Overall, the device presents a practical solution for improving swine management, promoting precision agriculture, and enhancing livestock welfare.