The lambda 3-plus (λ3+) Hyperspectral Data Analysis Studio
The λ3+ is Spectricon’s custom software platform designed for analyzing and processing spectral cube data acquired by MUSES9 cameras. It integrates modern machine learning tools to customize, simplify, and automate complex analytical tasks. The platform generates application-specific training files that can be directly integrated into the acquisition software module. Trained MUSES9 cameras produce material identification maps, which can be displayed in real time. The λ3+ studio offers intuitive user interfaces, online operational guidance, and can be easily integrated into machines and workflows.
The λ3+ basic functions
Spectral Cube Visualization
- Reviewing spectral cube images
- Pixel or pixel group-level spectroscopy
- Comparing spectral profiles
- Spectral clustering mapping
- ROI spectra statistics
Pre-processing
- Resampling of spectral data
- Resampling of spatial data (binning or cropping)
- Converting cube image format
- Noise suspension filtering
Spectral Image Processing & Analysis
- Brightness adjustment
- Contrast enhancement
- Image equalization
- Image Intensity histogram and threshold segmentation
RGB and False Color Image Analysis
- Color imaging from RGB data
- False color imaging from UV, VIS, IR data
- Full or ROI image color histogram analysis
- Imaging colorimetry indices (HSV, LAB, ΔΕ)
- Color index-based image segmentation
Spectral Image Calculations & Projections
- Phenology-based Vegetation index Classification (PVC)
- Phasor Clustering
- Principal Component Analysis (PCA)
- Linear Discriminant Analysis (LDA)
Spectral Similarity Measures
- Spectral Angle Mapper (SAM)
- Euclidean Distance (ED)
- Spectral Correlation Measure (SCM)
- Spectral Information Divergence (SID)
- Mean Squared Error Measures (MSEM)
Unsupervised Classification Algorithms
- K-means clustering
- K-nearest neighbors (KNN)
- ISODATA
- MeanShift
- DBSCAN
Machine Learning Toolbox
- Deep learning models: SVM, RF, CNN
- Customizable labelled spectra databases
- Generates application specific training files
- Training files are plugged in the acquisition SW for the real-time identification/shorting