Research On Technology Of Lunar Crater Detecting Based On Image Processing

In this paper, we combined with digital image processing, computer vision, patternrecognition and artificial intelligence to study pretreatment feature extraction andclassification methods of lunar crater image, and after that we made a lunar crater recognitionsystem based on image processing.We studied the processing methods crater image, and introduced image filtering and edgedetection processing methods, and improved the filtering algorithm, and introduced histogramspecification algorithm with good effects. Then, we deeply studied Haar-like features andPHOG features of crater image for particularity of crater image. The extracted crater imagefeature decided the accuracy of crater recognition, and extracting a good feature also cangreatly improve the recognition accuracy of the crater image, and it can shorten the calculatetime required by the system. and extracting a good feature also can greatly improve therecognition accuracy of the crater image, and it can shorten the calculate time required by thesystem. In this paper, we also studied recognition algorithms of craters image, an d introducedcommon template matching and Bayesian classification algorithm, we focused on theclassification algorithm based on AdaBoost and SVM. Particularly, it details the analysis andcomparison of the different characteristics with different classification algorithms andalgorithm combining the recognition accuracy and the time required, and researching thecomparative and analysis algorithms of AdaBoost classification combined with Haar-likefeatures, and SVM classification combined with PHOG features, and AdaBoost classificationcombined with SVM classification in this paper.In this paper, we designed a software package of automatic identification system basedon the methods and techniques of image processing and feature extraction and imageclassification, and verified these methods by the software package. Experiments show that thelunar craters system has faster recognition speed and higher accuracy, and it has somepractical significance to the identification of impact craters.