Comparison of Biodegradable and Non-Biodegradable Waste Detection Algorithm using YOLO v7 vs Faster R-CNN for Auto Segregation

This paper presents a detailed comparative analysis of biodegradable and non-biodegradable waste detection algorithms using YOLO v7 and Faster R-CNN, implemented for automated segregation on a Raspberry Pi platform integrated with a camera and a servo motor. The study aims to evaluate the performance of these state-of-the-art object detection models in terms of detection accuracy, processing speed, and computational efficiency in a real-time waste management system. YOLO v7, renowned for its rapid detection capabilities and lower computational demand, is compared against Faster R-CNN, which is recognized for its superior accuracy but higher computational cost. Experiments were conducted to assess the models' performance in identifying and classifying various waste types, and the results indicate a trade-off between speed and accuracy, with YOLO v7 providing faster detections suitable for real-time applications, while Faster R-CNN offers more precise detections at a slower pace. The integration of these algorithms with a servo motor facilitates accurate physical segregation of waste, showcasing the practical implications of each model's deployment in resource-constrained environments like the Raspberry Pi. This research highlights the potential and limitations of both algorithms, providing valuable insights for developing efficient and effective automated waste segregation systems.