Region-based image coding with multiple algorithms

The wide usage of small satellite imagery, especially its commercialization , makes data-based onboard compression not only meaningful but also necessa ry in order to solve the bottleneck between the huge volume of data generat ed onboard and the very limited downlink bandwidth. In this paper, we propo se a method that encodes different regions with different algorithms. We us e three shape-adaptive image compression algorithms as the candidates. The first one is a JPEG-based algorithm, the second one is based on the object- based wavelet transform (OWT) method proposed by [1], and the third adopts Hilbert scanning of the regions of interest followed by one-dimensional (1- D) wavelet transform. The three algorithms are also applied to the full ima ge so that we can compare their performance on a whole rectangular image. W e use eight Landsat Th I multispectral images and another 12 small satellit e single-band images as our data set. The results show that these compressi on algorithms have significantly different performance for different region s. For relatively smooth regions, e.g., regions that consist of a single ty pe of vegetation or water areas etc, the 1-D wavelet method is the best. Fo r highly textured regions, e.g., urban areas, mountain areas, and so on, th e modified OWT method wins over the others. For the whole image, OWT workin g at whole image mode, which is just an ordinary 2-D wavelet compression, i s the heat. Based on this, we propose a new data-based compression architec ture that extracts particular regions according to the application of inter est and then involves different algorithms to encode different regions in o rder to achieve better performance than traditional onboard compression sch emes in which a fixed compression method is applied to the whole image no m atter what the application is. This approach is most appropriate for use wi th images captured by microsatellites, which are commissioned for specific applications in which one knows a priori which class of region the user is interested in.