Freshness to Forecast: Smart Pricing and Quality Detection of Products via YOLOv8 and AI Regression

The agricultural commodity market in India suffers from fragmented supply chains, price fluctuations, and limited real-time quality assessment. Farmers often depend on intermediaries, leading to reduced profits, delayed sales, and significant post-harvest losses, particularly for perishable produce. Existing systems mainly target narrow tasks such as fruit counting or yield estimation, lacking comprehensive solutions that integrate quality evaluation with market intelligence. This research proposes a Commodity Intelligence System (CIS) that addresses these challenges by combining YOLOv8-based deep learning for accurate classification of multiple crops with regression models for freshness detection and market price prediction. The CIS is trained on a custom dataset of five widely consumed commodities—capsicum, onion, potato, apple, and banana—captured under diverse conditions to enhance model robustness. Experimental results show improved classification accuracy and predictive performance over conventional approaches. The proposed CIS has potential applications in smart agriculture and supply chain intelligence, enabling transparent pricing, reduced wastage, and fairer market access for farmers and retailers.