Research On The Navigation Orientation Object Extraction Of Remote Sensing Images

Navigation orientation object which is an obvious fixed feature selected to determined orientation or indication targets in ship sailing, is an important part of the coastal zone. Aiming at the status that the relatively weak ability to obtain geospatial information of coastal zone, this thesis mainly focus on how to extract navigation orientation object (tanks, wharf and breakwater) quickly and efficiently from remote sensing images.The main work and innovation of this thesis include:(1) Based on the analysis of the current situation of the development of high spatial resolution satellite imagery, and the obviously insufficient extraction method of sailing destination feature in mapping, such as coastal topographic map, it defines the concept of navigation orientation object and expounds its background and significance. Meanwhile, this thesis analyzes the problems and difficulties faced in the process of extraction at home and abroad.(2) Based on the analysis of the related characteristics of remote sensing image object, it introduces the basic theory of Hough transform circle detection aimed at inshore navigation orientation object (tanks). Meanwhile, in the theory of image segmentation and contour extraction of navigation orientation object, Snake model and methods are discussed.(3) Aimed at the problem that the low efficiency of traditional Hough transform in tank target extraction, it puts forward a fusion method of extracting tank in remote sensing image based on the visual attention mechanism and the Hough transform. On the basis of the analysis of image significant characteristics and oil tank significant areas, it mainly introduces the calculation method of tank target significant area based on visual attention mechanism. Use the mechanism to get the oil tank target significant area, then realize the extraction of tank target by Hough transform circle detection of the significant area.Test results show that the method has high extraction efficiency, accuracy and reliable characteristics.(4) It analyses the image characteristics of wharf and breakwater,and proposes a method based on the multimodal gray-level histogram threshold segmentation to extract them. It improves the extraction method based on Snake model, and uses extracting test to analysis of the reliability of the two extraction methods in the thesis.