Machine learning-driven spectroscopy for metal identification in waste

Overview

Eco Mercantile presents an advanced solution combining machine learning with spectroscopy to identify metals in waste rapidly. This technology leverages intellectual property from Naval labs to design a laser diode system capable of varying energy intensity and quantity. By integrating this with spectrometers, the system processes lightwave data to generate signals for machine learning algorithms, facilitating quick and accurate metal identification. This innovative approach aims to create a 3D characterization of materials, enhancing waste sorting and recycling efficiency.

Technical specifications

• Laser Diode System: Designed to generate varying laser energies for optimized material scanning.
• Spectrometer Integration: Processes lightwave data for material analysis.
• 3D Characterization: Utilizes additional sensors to develop a comprehensive material profile.
• Machine Learning Algorithms: Employs advanced neural networks to analyze spectral data and map elemental distribution.
• Optimization Steps:
  • Laser diode testing and absorption spectra measurement.
  • Beam collimator optimization for precise focus control.
  • Spectrograph calibration for enhanced resolution and sensitivity.
  • CMOS sensor and ADC optimization to improve signal-to-noise ratio.

Technology readiness level

Currently at Technology Readiness Level 2, this solution is under development with ongoing negotiations for exclusive field licenses. Supported by a lab letter, it is poised for further validation and proof of concept development, focusing on rapid metal identification in waste.