MUMBAI, India, Sept. 22 -- Intellectual Property India has published a patent application (202541023065 A) filed by Arun Kumar U B; Mathavan K; Kaviarasu Venkatesan; Sujith P; and Reshmi T S on March 14, 2025, for Cnn-Based Grading Of Pithecellobium Dulce Ripeness Using Thermal Imaging.

Inventors include Arun Kumar U B; Mathavan K; Kaviarasu Venkatesan; Sujith P; and Reshmi T S.

The application for the patent was published on September 18, 2026, under issue no. 38/2026.

Abstract: This invention presents an advanced CNN-based system for grading the ripeness of Pithecellobium dulce usmg thermal imaging technology. Traditional methods, like manual inspection and destructive. testing, often lead to inaccuracies, inefficiency, and post-harvest losses. By integrating thermal imaging with Convolutional Neural Networks (CNNs), this invention overcomes these limitations through non-invasive, data-driven classification. The system works by capturing thermal omages of the fruit, detecting subtle temperature variations associated with metabolic heat produced during different ripeness stages. These images undergo preprocessing to enhance key features, followccl by CNN-based analysis that extracts temperature patterns nnd classifies the fruit into ripeness categories - unripe, semi-ripe, or fully ripe. This ensures accurate, consistent results, unaffected by external factors like surface irregularities or lighting. The invention offers a scalable, real-time solution for agricultural environments, supporting faster sorting, reducing waste, and ensuring higher market quality. Additionally, the CNN model continuously improves by learning from new data, making the system adaptive to different fruit batches and conditions. While designed for Pithecellobium dulce, the methodology can be adapted for other fruits with similar ripening characteristics, demonstrating its versatility. This novel approach transforms fruit grading, enhancing efficiency, minimizing losses, and providing a sustainable, cost-effective alternative to conventional ripeness assessment methods.

Disclaimer: Curated by HT Syndication.