Research And Implement Of Harbor Detection In Remote Sensing Image

As an important military and civil building, the harbor is a target including much information. So it become an important mission in the military field to research the harbor recognition in the remote sensing images, and it is also one of the hotspots in Pattern Recognition and Image Processing. The harbor recognition technology has an important application foreground. It is also a very complicated problem in the area of Pattern Recognition for the sake of the complex environment and features of the harbor.Firstly, the characteristic of the harbor is analyzed in this paper and some distinct feature is concluded. Then, after the prophase processing, a detecting method is proposed in the paper to recognize the harbor in the image. And a new edge detection method is also proposed in this paper: An Edge Detection Method Based on Maximal Entropy and Morphology. It is proved by the experiment that the method proposed in this paper can detect the harbor in the image efficiently, and it is immune from the rotation and the noise. Through the edge detection method in the paper, the edge of gray image is detected efficiently and it is single-pixel line. It remains the image information at most while picks the edge quickly and precisely。The main contributions given in this dissertation as follows:(1) Based on analyzing the feature owned by the harbor target, conclude the feature between the harbor and the water area. Then detect the harbor in the image through horizontal and portrait scan. The experiment shows that the method is immune from the rotation and the noise, because of the blur process and stain elimination.(2) A new edge detection method is proposed, and he experiment indicates that it can obviously improve the quality of edge detection in this method compared with traditional methods, and the detected edges are all single-pixel line. It remains the image information at most while picks the single-pixel edge quickly and precisely。...