Abstract:Recent advancements in computer vision, particularly in detection, segmentation, and classification, have significantly impacted various domains. However, these advancements are tied to RGB-based systems, which are insufficient for applications in industries like waste sorting, pharmaceuticals, and defense, where advanced object characterization beyond shape or color is necessary. Hyperspectral (HS) imaging, capturing both spectral and spatial information, addresses these limitations and offers advantages over conventional technologies such as X-ray fluorescence and Raman spectroscopy, particularly in terms of speed, cost, and safety. This study evaluates the potential of combining HS imaging with deep learning for material characterization. The research involves: i) designing an experimental setup with HS camera, conveyor, and controlled lighting; ii) generating a multi-object dataset of various plastics (HDPE, PET, PP, PS) with semi-automated mask generation and Raman spectroscopy-based labeling; and iii) developing a deep learning model trained on HS images for pixel-level material classification. The model achieved 99.94\% classification accuracy, demonstrating robustness in color, size, and shape invariance, and effectively handling material overlap. Limitations, such as challenges with black objects, are also discussed. Extending computer vision beyond RGB to HS imaging proves feasible, overcoming major limitations of traditional methods and showing strong potential for future applications.

A PCA-Based Material Color Classification Method Via LED Sensing

A Pca Based Automatic Image Categorization Approach Using Dominant Color Features

Fast and High-Accuracy Spectral Measurements of LED by Linear CCD Sensor and Software Calibration

Discriminative illumination: per-pixel classification of raw materials based on optimal projections of spectral BRDF

LED-based color sensing and control

Waste Material Classification: A Short-Wave Infrared Discrete-Light-Source Approach Based on Light-Emitting Diodes

Wavelength Dependence of Colorimetric Properties of Lighting Sources Based on Multi-Color LEDs

Optimized principal component analysis for camera spectral sensitivity estimation.

Color Sensing and Illumination with LED Lamps

Luminescence mechanism analysis on high power tunable color temperature Chip-on-Board white LED modules

Multi-chip Dynamic White Light Emitting Diode with High Level Photobiological Safety and Good Color Fidelity

Materials Identification of Polarized Pulse Laser Detection Based on Sparse Autoencoder and Softmax Classifier Framework

Spectral Optimization Based Simultaneously on Color-Rendering Index and Color Quality Scale for White LED Illumination

Skin Lighting Study Based On Multispectral Images

A Deep Learning Approach for Pixel-level Material Classification via Hyperspectral Imaging

Optimization of Illumination Performance of Trichromatic White Light-Emitting Diode and Characterization of Its Modulation Bandwidth for Communication Applications

Study on a system for LED's photometric and colorimetric measurement based on a multi-channel spectrometer

Waste Material Classification Based on a Wavelength-Sensitive Ge-on-Si Photodetector

Multispectral Imaging Based on a Multichannel LED System and RGB Camera

Remote Control and Noninvasive Detection Enabled by a High-performance NIR pc-LED

Filterless Photosensors with Intrinsic Color Perception and Flexibility