- 期刊
- OpenAccess
Complex Wavelet Package Image Restoration to Support Operational Application
並列摘要
Operational application image restoration method with high quality, high efficiency and robustness is presented, for high resolution optical remote sensing satellite. Firstly, images are block segmented according to CCD characteristics, and periodic plus smooth decomposition and dequantization are processed. Then, noise and MTF curves are estimated adaptively using spectrum masking for taking into account the on-orbit dynamical imaging properties. Thirdly, deconvolution filters are optimized in different frequency in order to balance among the blurring, aliasing, noise and ringing effects, based on known noise and MTF's shape and value. Fourthly, adaptive threshold to color noise are treated rapidly utilizing the shift invariance and the direction selectivity of complex wavelet package (CWP) unique features. Fifthly the weak texture regions are reserved. Experiments shows that the method can significantly improve image sharpness, with universality adaption, and has a good operational application prospects.