Waste Segregation using IoT and Machine Learning

India is one of the most populated countries in the world, and its population is escalating rapidly, being very close to surpassing China in the near future. As the population increases, the waste generated is also rising rapidly. The per-capita waste generation in India is approximately 0.5 to 0.7 kg per day, which can lead to significant environmental setbacks. To tackle this issue, the Indian government has initiated several waste management programs that have demonstrated significant advancements. Nevertheless, certain segments still lack the intended impact on the populace and exhibit inconsistencies. To bolster these efforts, an automated waste segregation system has been introduced. This paper focuses on developing an automated waste segregation model to make the waste management system more efficient and automated. By achieving this, we aim to help clean our environment and invest our time in working towards a circular economy as a society. We employed YOLOv5 technology in conjunction with the Roboflow dataset to classify waste into four fundamental categories: Dry waste, Wet waste, E-waste, and Plastic waste. The model has been trained on 31,690 images and has achieved a high classification rate. The Roboflow dataset was used to develop a more accurate, flexible, and scalable waste segregation model, paving the way for effective waste segregation. The proposed model has achieved a mean CV of about 97%, demonstrating its potential to effectively segregate real-world waste into various categories. The efficacy of this waste segregation model lies in automating the waste segregation system, which not only reduces human labor in disease-prone situations but also improves overall efficiency.