Real-Time Detection of Domestic Waste Based on Deep Learning

This study presents a waste classification target detection network based on the RT-DETR model, demonstrating the potential application of deep learning technology in environmental protection. The network accurately identifies and classifies various items in household waste. As an end-to-end network for real-time monitoring, RT-DETR is relatively simple in application deployment and allows for easy integration of different network modules for targeted improvements. This study not only verifies the effectiveness of the RT-DETR model in real-time monitoring of waste classification but also provides new ideas and methods for future research in environmental protection. The detection network achieved an mAP of 66.75% and an FPS of 45, indicating that the RT-DETR model has high accuracy and robustness while maintaining real-time processing capabilities, effectively enabling automatic classification of household waste.