| This study presents a framework for the study of the restoration of degraded and corrupted NOAA-AVHRR (Advanced-Very High-Resolution Radiometer) images, using wavelets decomposition. This framework consists of: image gathering system, resampling in changes of format, wavelets, subband decomposition, and selection of detail wavelet coefficients and filtering to perform image restoration.;A Landsat-MSS (Multispectral Scanner) scene was used to simulate the NOAA-AVHRR gathering system. The system model considered the initial scene, the distortion, and the additive gaussian white noise. The output was the high-resolution data format (Local Area Coverage). A resampling simulated the format change, resulting in the low resolution data format (Global Area Coverage). The restoration was done using multiresolution by wavelet decomposition and the Donoho and Johnstone thresholding criteria. Two types of thresholding (hard and soft) were applied to the wavelet coefficients.;The methodology based on wavelet decomposition employed in this study for the restoration of NOAA-AVHRR images produced a significant increase in the SNR (Signal to Noise Ratio) over other methodologies. |
