The increase in population is giving rise to an exponential increase in the amount of waste that is being generated in day-to-day life. The inappropriate dumping and disposal of these wastes result in major damage to the air, land, and soil by affecting the quality, mineral content, productivity, and nature. In order to carry out efficient waste management, the recyclable wastes should be treated properly. The first step toward recycling is the segregation of wastes into their respective class. The project mainly focuses on the detection and classification of waste materials using the You Only Learn One Representation (YOLOR) object detection algorithm and image processing techniques. A custom dataset has been prepared, and the number of classes of wastes has been increased to 10. The custom dataset has the following classes of waste: cardboard, electronic waste, glass, kitchen waste, medical waste, metal, paper, plastic waste, textile waste, and trash. The classification is performed by the YOLOR algorithm over the custom dataset prepared with a number of preprocessing steps. The detection performance is compared with the previous versions of the You Only Look Once (YOLO) algorithm, such as YOLOV4, YOLOV5, both with and without the preprocessing steps, to prove the effect of preprocessing. The model is analysed in terms of Mean Average Precision (mAP), box loss, class loss, and object loss. The proposed model achieves a Precision of 94.2%, a Recall of 93.3%, and an mAP of 95.9%. The proposed model shows a box loss of 0.025, a class loss of 0.0001, and an object loss of 0.069.