Content-based onboard compression for remote sensing images

New-generation instruments on spacecraft are collecting a large amount of information at an increasing rate, which makes the onboard data compression a challenging task. Moreover, existing compression methods usually scan an image in a fixed way without considering the content of the image, which makes the performance improvements of these methods often marginal at best. In this paper, we present a novel, content-based, adaptive scanning (CAS) scheme for onboard compression. For a remote sensing image, first, the wavelet transform is performed. Second, an adaptive scanning method is proposed, which can provide different scanning orders among and within subbands, respectively. The former aims at organizing the codestream according to the importance of subbands, and the latter focuses on preserving the texture information as much as possible. Finally, the binary tree codec is utilized to code the 1-D coefficient array after scanning. Experimental results demonstrate that compared with other scan-based compression methods, including CCSDS, JPEG2000, and even the state-of-the-art adaptive binary tree coding (BTCA), the proposed compression method can effectively improve the coding performance. In addition, the method does not use entropy coding or any complicated components, which makes it extremely suitable for onboard compression.