A survey on image completion techniques in remote sensing images

Abstract

Aèstract-Nowadays, reconstruction of satellite images is one of the important challenges in technology. Sometimes, remote sensing images have dead pixels or pixel missing gives poor visual quality. It is mainly due to the presence of clouds, fogs, or shadows since these images are acquired by sensors at different seasons. The other serious problems in remote sensing images are instrumentation error, registration error and losses of image data during transmission. To get a good visual quality images, degraded remote sensing images are processed by using the application of inpainting. The main objectives of this image inpainting approaches are to fill lost parts of images, delete unwanted objects, remove noise in images and to enhance images quality. Image completion or image inpainting process is one in which the damaged portions are reconstructed or to fill the lost regions using data collected from surrounding areas in original image. So far, a number of inpainting algorithms are available. This paper gives a detailed survey of some inpainting methods which are suitable for remotely sensed images.