The Analysis Of Lunar Surface Features Based On Remote Sensing Image

It is an international popular topic to explore lunar. It is not only the front of science realm, but also the body of the comprehensive national strength. Along with the successful blast-off and return of"Shenzhouâ„–5"and"Shenzhouâ„–6", our country has taken the lunar exploration into the national develop agenda, and launch on the detailed development plan of lunar exploration by three-steps. With the technical development of the spaceflight and remote sensing especially the secondary planet photograph measure, it has become the key to explore the lunar with the technique of lunar exploring secondary planet photograph measure. This article set about researching the disposal and analysis of planet remote sensing photo based on diagnostic of lunar surface. The major aim is to research the theory method for analyze the image of lunar surface after explored lunar.This article explains the basic theory of disposing the image remote sensing at first, and then distilled the necessary feature successfully with the series method of image strain, linear measure, image division, feature extraction based on MATLAB.First, this article introduced the pretreatment method of image"a"and"b". In the part of image filtering, mainly introduced contrast boost up filtering and smoothness filtering. Particularly in choose a method of smoothness filtering, It has been found out in this article that the most suitable method of the image that this text adopt by compare three methods (mean filtering, median filtering, adaptation wiener filtering); in the part of image operation and region disposal, mainly introduced the methods of deal with the pixel of image, including image inverse operation and image subtraction operation. And it describes the method of dealing with image region with mask-template in the same time; in the part of image enhancement mainly introduced the method of histogram equalization based on gray scale transformation in image. All methods that the article used have the homologous picture to carry on the comparison, and carry on the analytical elucidation to the contrast result diagram of the image, enabling the reader more clearly to the text withprocessing result of method to image, thus prove this thesis to having obvious function to the processing of the picture in the method adopt, make the image obtain the better result in follow-up handle.In the part of image segmentation, there are two kinds of methods had been used, and described them in detail. In threshold methods for image segmentation, the article gets the conclusion that segmenting image by the method of OTSU through the comparative analysis of traditional method, such as double peak method, iteration method and OTSU. The processing result that passes the image explains that the method is fit for the image of this article. In the part of edge detection arithmetic demonstration, the article gets the conclusion that Canny operator excel the others traditional operator (Robert operator,Sobel operator,Prewitt operator,Log operator) through the comparative analysis of these operators, and the processing result that passes the image explains that the method is the best method for the image of this article.The text still introduces mathematical morphology application to image processing, the method of image recognition process and the method or the basis based on feature of parts in image. First, the article introduces mathematical morphology and its application, and then introduces morphological structural element, basic operations and the application in dealing process for image. Morphology has important function and it has become a key role in dealing process for image by binary image closure operation, small area remove, connected region label, feature extraction and so on. The article described the theory and the method of image recognition, emphasized to elaborate the method and homologous basics that the characteristic classification that this text use identify, carried on the argument and analysis of the accurate degree in recognition. The experimental results show that the methods are accurate and feasible, but existing faults and limitations in some extent. The last part of article is a generalization part about all methods of article. It described entirety flow chart of method in characteristic classification and recognition about image"a"and"b". Then inextenso and vividly denote dealing process of image by marginal data, which make reader know the whole process of deal with image very clearly; the text use two ways to deal with another polarized light image for demonstrate that two methods is feasibility, and analyze the advantages and the disadvantages of two methods base on treatment result.Lastly, the article carried on with the error and the limit that the methods mayproduce overall of analysis and dissection. The image that this paper used most come from remote sensing pictures apart from the moon about between 1,200,000 meters to 2,600,000 meters on SMARTâ… which the ESA(European Space Agency) launch on, So the images were affected easily. Some essential factors such us the radiation angle of sun,the distance of remote,the definition of image and the types of sensors have great effect for the method that this paper used; There are still some errors exist in methods of image process. In prepare judge process of the picture characteristic classification, the method of judge with the center mean gradation and the threshold of image segmentation is not omnipotence because it only has the effect that we want in three images; the images that this article used all about of the lunar surface, so there are great limit in image origin and this limit can make great confine in method argumentationThe main aim to write this article is to explicit way of thinking, corroborating the method, pointing out the shortage of method in this article, and also hope afterwards can go forward to go the complement and developments in this foundation. So this text also has the certain theories value and actual application value.