Abstract:Hyperspectral compression is one of the most common techniques in hyperspectral image processing. Most recent learned image compression methods have exhibited excellent rate-distortion performance for natural images, but they have not been fully explored for hyperspectral compression tasks. In this paper, we propose a trainable network architecture for hyperspectral compression tasks, which not only considers the anisotropic characteristic of hyperspectral images but also embeds an accurate entropy model using the non-Gaussian prior knowledge of hyperspectral images and nonlinear transform. Specifically, we first design a spatial-spectral block, involving a spatial net and a spectral net as the base components of the core autoencoder, which is more consistent with the anisotropic hyperspectral cubes than the existing compression methods based on deep learning. Then, we design a Student's T hyperprior that merges the statistics of the latents and the side information concepts into a unified neural network to provide an accurate entropy model used for entropy coding. This not only remarkably enhances the flexibility of the entropy model by adjusting various values of the degree of freedom, but also leads to a superior rate-distortion performance. The results illustrate that the proposed compression scheme supersedes the Gaussian hyperprior universally for virtually all learned natural image codecs and the optimal linear transform coding methods for hyperspectral compression. Specifically, the proposed method provides a 1.51% to 59.95% average increase in peak signal-to-noise ratio, a 0.17% to 18.17% average increase in the structural similarity index metric and a 6.15% to 64.60% average reduction in spectral angle mapping over three public hyperspectral datasets compared to the Gaussian hyperprior and the optimal linear transform coding methods.

Hyperspectral image lossless compression algorithm based on adaptive band regrouping

Multispectral image compression by cluster-adaptive subspace representation

Compressive Single-Pixel Hyperspectral Imaging Using RGB Sensors

Distributed Lossless Coding Techniques for Hyperspectral Images

Distributed lossless coding of hyperspectral images

Hyperspectral image recovery based on fusion of coded aperture snapshot spectral imaging and RGB images by guided filtering

Spatial-spectral Encoding and Dictionary Optimization in Compressive Single-Pixel Hyperspectral Imaging Based on Mutual Coherence Minimization

Lossy Compression Algorithm of Multispectral Remote Sensing Images Based on IHS Transform

Band Selection for Hyperspectral Imagery Using Affinity Propagation

A Compression Algorithm of Hyperspectral Remote Sensing Image Based on Lifting Scheme

Lossless hyperspectral image compression by combining the spectral decorrelation techniques with transform coding methods

A Robust Band Compression Technique for Hyperspectral Image Classification

Probability Prediction Network With Checkerboard Prior for Lossless Remote Sensing Image Compression

Three-stages Hyperspectral Image Compression Sensing with Band Selection

Context-Based Lossless Compression of Mosaic Image with Bayer Pattern

Fast Multi-level Clustering Lossless Compression Algorithm for Remotely Sensed Images

A Compression Algorithm of Hyperspectral Remote Sensing Image Based on Vector Quantization

Lossless and Near-Lossless Compression Algorithms for Remotely Sensed Hyperspectral Images

LOSSLESS INTER-ARRAY PREDICTIVE CODING FOR SUBPIXEL-SHIFTED SATELLITE IMAGES BASED ON TEXTURE ANALYSIS

Learned Hyperspectral Compression Using a Student's T Hyperprior

Hyperspectral Image Compression Via Cross-Channel Contrastive Learning.