Study Runway Recognition Based On Airport Remote Sensing Image

This thesis has proposed algorithms on automatic extracting and recognizing airfield runway from remote sensing images. As an important part of the technology of remote sensing images recognition, a great many scholars have exerted themselves. In accordance with the needs of this task, methods in this thesis are improved based on former research results and combining these methods effectively, we get satisfying results.The work including:(1) Analyzing classical digital image enhancement algorithm in the remote sensing image, the fuzzy enhancement algorithm is used to preprocess the remote sensing images. Based on this foundation, it puts forward an improved fuzzy enhancement algorithm. This algorithm raises the speed of algorithm handling on the premises of not to affect the handling quality of image.(2) Extracting airport areas:first, the method of threshold division is applied to image segmentation, and then the pixel labeled algorithm is used to extract the maximum cross-connection domain. And eventually the improved algorithm of ROI is applied to implement the orientation of the possible airport areas.(3) Using Canny Operator to extract the edge of the image on the basis of the characteristic of remote sensing image and the compartment of several classical edge segmentation algorithms. Based on analyzing the classical Canny Operator, proposing an improved Canny algorithm on the basis of the multi-scale morphological gradient operation.(4) The line extraction technology based on Hough transformation has been mainly studied. A beeline extraction technology based on random Hough transformation was carried out after the analyzes of the classical Hough transformation, random Hough transformation, CGHT. At last, the thesis recognizes the area of runway in terms of the intrinsic feature of the airfield runway.