A thermal noise suppression method for increasing lossless compression ratio of the remote sensing images

In order to improve data transmission efficiency between ground and satellite, image compression methods have been implemented onboard for years. However, the thermal noise generated by charge-coupled device (CCD) will decrease the compression ratio of the onboard image data and there is few works to be undertaken on noise suppression oriented for increasing image compression ratio. Therefore, a Homogeneity and Insulation based Forward-And-Backward Anisotropic Diffusion (HIFABAD) method is proposed for increasing lossless compression ratio of the remote sensing images. Different from anisotropic diffusion methods which aim for over-smoothing the image, a novel noise estimation method based on local homogeneity measurement and insulated boundary extraction is proposed for terminating the iteration of HIFABAD. Furthermore, based on the result of homogeneity measurement, the proposed method adds an adaptive parameter for noise suppression of the images with different degrees of homogeneity. To assess efficiency of the proposed method, multispectral and panchromatic image data of the Beijing-1 small satellite are selected to be the data sources. The experimental results show that a higher lossless compression ratio and an acceptable quality of the denoised image can be obtained based the proposed HIFABAD method compared with other existing anisotropic diffusion methods.