[Lossless compression of hyperspectral image for space-borne application]

Area of Science:

• Remote Sensing
• Image Processing
• Data Compression

Context:

• Hyperspectral imaging generates large datasets, posing challenges for storage and transmission in space-borne applications.
• Existing lossless compression algorithms often struggle with hardware implementation complexity and efficiency.
• Wavelet-based compression schemes can be computationally intensive and difficult to implement on hardware.

Purpose:

• To develop a hyperspectral image lossless compression algorithm optimized for space-borne applications.
• To address limitations in hardware implementation, compression ratio, and processing time of existing methods.
• To enhance the efficiency and practicality of hyperspectral data compression for remote sensing.

Summary:

• A novel hyperspectral image lossless compression algorithm is proposed, utilizing intra-band prediction for the first spectral image and a two-step, bidirectional inter-band prediction for subsequent images.
• The inter-band prediction incorporates a bidirectional, second-order predictor and an improved Look-Up Table (LUT) prediction algorithm for enhanced prediction accuracy.
• Verification experiments using dedicated test equipment demonstrated the algorithm's fast and stable operation, achieving an average compression ratio of 3.05 bits per pixel (bpp).

Impact:

• The proposed algorithm significantly improves the average compression ratio by 0.14–2.94 bpp compared to traditional approaches.
• It effectively enhances lossless compression performance for hyperspectral imagery.
• The method successfully overcomes the hardware implementation challenges associated with complex compression schemes, making it suitable for space-borne systems.