BTC-Net V2: Spatial-Priority Hierarchical Reconstruction for Hyperspectral Image Compression (opens in new tab)
In recent years, lossy compression of hyperspectral images (HSIs) for spaceborne applications has garnered significant attention. The limited storage capacity and constrained transmission bandwidth of spaceborne equipment make it challenging to simultaneously balance the compression rate of transmitted data and the quality of locally reconstructed images. To address this challenge, we propose bit-level tensor data compression network (BTC-Net) V2—an advanced iteration of BTC-Net—that substant...
Read the original article