A Systematic Review of AI-Based Techniques for Automated Waste Classification

Area of Science:

• Environmental Science
• Computer Science
• Artificial Intelligence

Background:

• Traditional waste classification is labor-intensive and inefficient.
• Automating waste classification is crucial for effective waste management.
• Machine learning (ML) and deep learning (DL) present viable computational alternatives.

Purpose of the Study:

• To systematically review the application of artificial intelligence (AI), ML, and DL in automating waste classification.
• To analyze existing waste classification datasets and identify their limitations.
• To propose a roadmap for future research and development in AI-powered waste classification.

Main Methods:

• Conducted a systematic literature review (SLR) following PRISMA and Kitchenham guidelines.
• Analyzed over 97 studies on AI-based waste classification.
• Categorized AI techniques into ML-based, DL-based, and hybrid models.
• Reviewed over 15 publicly available waste classification datasets.

Main Results:

• Deep learning and hybrid AI models are dominant in current waste classification research.
• Convolutional Neural Network (CNN) architectures and transfer learning show significant promise.
• Key dataset limitations include imbalance, real-world variability, and lack of standardization.
• AI techniques offer a path towards efficient and automated waste management.

Conclusions:

• AI, particularly DL and hybrid approaches, is transforming waste classification.
• Addressing dataset limitations and focusing on practical deployment are key for future advancements.
• A structured roadmap prioritizes challenges and opportunities for AI in waste management, balancing accuracy, efficiency, and sustainability.