卫星应用中基于视觉显著性的感兴趣区域图像编码算法

Region-of-interest Image Coding for Satellite Application Based on Saliency of Visual Attention

为了满足遥感图像在卫星应用中有限信道下高保真传输的需求,本文提出了一种基于视觉显著性的感兴趣区域图像编码算法。首先,基于CCSDS标准的编码框架可以降低编码算法复杂度,便于实时实现,适合于卫星应用场合。其次,引入视觉注意机制,采用一种基于视觉显著性的自动提取算法提取感兴趣区域。最后,通过使用掩模来确定ROI区域,可以对任意形状的ROI区域进行编码。实验结果表明,与JPEG2000中的感兴趣区域图像编码算法相比,本文算法结构简单,便于实现;拥有优秀的主观视觉质量,符合人眼的感知,在高压缩比下能够为用户提供更多的信息。另外,本文算法可以作为CCSDS标准中的感兴趣区域编码算法。

To satisfy the demand for high-fidelity remote sensing image transmission in the limited channel, a Region-of-interest （ROI） image coding algorithm based on the saliency of visual attention is proposed. Firstly, the CCSDS coding framework can reduce the computing complexity, which is suitable for real-time applications （such as satellite application）. Secondly, an automatic ROI extraction algorithm, which is based on the saliency of visual attention, is proposed. Finally, the ROI is determined by using the mask and can be used for arbitrary shape. The experimental results have demonstrated that, compared with the algorithm in JPEG2000, the proposed algorithm is easy to be implemented. Moreover, it has good subjective visual quality, which is consistent with the human visual perception. Meanwhile, the proposed algorithm could be a ROI algorithm used for CCSDS standard.