<![CDATA[Laboratories serve as crucial facilities supporting practical activities, with a recommended maximum of 20 students, necessitating periodic monitoring to count the dynamic number of students within. The system utilizes the COCO dataset labeled ”person,” involving an approach with entry and exit preference lines, ID identification implementation, and object detection models YOLO v3 Tiny and Faster R-CNN ResNet50. The main system components, Raspberry Pi 3 Model B+, Raspberry Pi Camera 5 MP (f/1.3), and Raspberry Pi 7-inch Touch Display, are integrated for processing, real-time video recording, and image display functions. Test and evaluation results reveal that YOLO v3 Tiny achieves an 88.24% accuracy for entry counting and 75% for entry-exit counting, with an average processing rate of 4.89 FPS, while Faster R-CNN ResNet50 demonstrates lower accuracy, reaching 70.59% and 45.83%, with an average processing rate of 0.58 FPS.]]>
Copyrights © 2024
