Research On Remote Image Compression Structure And Implementation Of Important Circuit Modules

With the development of sensor technology, acquired remote sensing images havehigher resolution, resulting in rapid growth of image data. Therefore, the effectivereal-time compression of the data, and information transfer in limited time andbandwidth, has become an urgent problem to be solved. Based on the direction ofdiversity of remote sensing images, this paper introduces a high performance imagecompression encoder. First let the image be three coding modes, which are I4, I8andI16. And then every coding mode does the steps as follows: direction prediction,optimal transform, quantization, inverse quantization, inverse transform, reconstruction,rate-distortion optimization and entropy coding.Based on the analysis of the image compression encoder’s frame, this paper let thethree coding modes be parallel branches for processing. The I4/I8coding mode isdivided into nine-stage pipeline in multi-cycle, and the I16coding mode is processed bymultiplexing circuit with the serial structure. This paper also completes the design ofeach coding mode and gives the detailed data flow diagram. In addition, the importantcircuit modules are designed and implemented in FPGA, including non-separable K-Ltransform in the I4/I8mode, separable K-L transform in the I16mode, quantization,rate-distortion optimization and optimal information storage. Finally, this articlepresents the RTL-level simulation maps of each module and the integrated results of thisremote image compression encoder.