Ai-based vision system for cocoa bean fraction identification

Overview

UHV Technologies presents an AI-based vision system designed to revolutionize cocoa processing by accurately identifying and classifying cocoa nibs, shells, and fines. This innovative solution leverages our extensive expertise in AI/ML-based sorting technologies to improve component separation and yield. By utilizing visible cameras and advanced AI models, our system promises over 95% accuracy in diverse conditions, significantly reducing contamination and losses compared to traditional sorting methods.

Technical specifications

• AI-driven image analysis: Utilizes shape, size, texture, and color to distinguish cocoa components.
• High accuracy: Achieves over 95% identification accuracy in varying processing conditions.
• Proven technology: Based on patented vision-based AI technology with a successful track record in sorting challenging material streams.
• Real-time processing: Capable of real-time monitoring and dynamic control, adapting to conveyor speeds for industrial-scale throughput.
• Integration options: Can enhance existing sorting processes or integrate with custom sorting systems, including air jets.

Technology readiness level

Currently at TRL 3, this system is in the development phase with a focus on capturing 10,000 images for model training, optimizing algorithms for high throughput, and determining effective integration strategies. Future validation plans include training and testing multiple CNN models to select the optimal one for industrial application.