Research On Remote Sensing Digital Image Mosaic Algorithm

Currently, the technology of remote sensing has been widely used in military,meteorological, environmental and agricultural fields. And At the same time, the technology of remote sensing image mosaic has been into the limelight in the field of computer science and image science, the technique of remote sensing image mosaic is major research areas in the development of remote sensing technology, its purpose is that different pieces of the original remote sensing image is mosaicked into a large remote sensing image, so that people make full use of the remote sensing data. This paper mainly studies the algorithm of remote sensing image mosaic, the completion of the specific work is as follows:(1) This paper introduces the common algorithm in the field of remote sensing image, and makes a detailed study on the fusion of the overlapped area and the morphology of the image. For example, a mosaic algorithm of the wavelet transform method, Laplacian pyramid, image difference in the overlapping area and effective area of angle was studied, and the effect of the four algorithms are further analyzed and discussed.(2) In order to solve the problems that image’s entropy of information decline obviously and boundary line phenomenon appear obviously in the processing of aerial remote sensing image’s mosaic, this paper proposes wavelet image fusion based on structure similarity and use the structural similarity as the selection standard for the fusion rule. Experimental simulation results show that, the mosaic image has a high information entropy, high standard deviation and high-definition quality. The weighted average method is used to deal with the joint seam, the obvious phenomenon of seam line has been greatly improved, and the processing speed is relatively fast than the relative classical algorithm.(3) According to the obvious phenomenon of remote sensing image mosaic seam line, This paper presents a method of remote sensing image mosaic based on watershed image segmentation technology, using watershed image segmentationalgorithm to extract the linear object in the overlapping area of remote sensing image and putting this line as the seam line. This method has low computational complexity and high efficiency, and can effectively avoid the obvious phenomenon of the seam,and it is difficult to find the existence of the seam line in the subjective vision.