Compressive hyperspectral image classification using a 3D coded convolutional neural network.

Area of Science:

• Remote Sensing
• Computer Vision
• Optics

Background:

• Hyperspectral image classification (HIC) faces challenges due to large data volumes.
• Existing methods often require full data cube reconstruction, increasing processing demands.

Purpose of the Study:

• To develop a novel deep learning approach for HIC using compressive measurements.
• To improve HIC efficiency and accuracy by avoiding full data cube reconstruction.

Main Methods:

• A 3D coded convolutional neural network (3D-CCNN) was developed.
• The coded aperture in CASSI was integrated as a network layer.
• An end-to-end training optimized network parameters and coded apertures simultaneously.

Main Results:

• The proposed 3D-CCNN effectively classifies hyperspectral images using compressive measurements.
• Joint optimization of network and coded apertures significantly improved classification accuracy.
• The method demonstrated superior performance compared to state-of-the-art HIC techniques.

Conclusions:

• The novel deep learning strategy offers an efficient solution for HIC with CASSI systems.
• Synergy between deep learning and coded apertures enhances classification performance.
• This approach mitigates challenges associated with large hyperspectral data cubes.