Based On The Characteristics Of Remote Sensing Image Segmentation To Identify The Prototype Design And Development Of Technology And Software Systems

This paper takes the high spacial-differentiating remote sensing image material as research objects, studies and realizes four feature-oriented image segmentation methods: orient phase method, fuzzy threshold with genetic algorithms method, Gaussian Markov Random Field Model texture segmentation method and moment feature-based texture segmentation, analyses their fundamental theories, their merits and drawbacks in detail, and validates them by some experiments. Furthermore, it designs and develops a corresponding software system model, including the system interface, the system architecture and the system functional modules, then analyzes the integration technology of remote sensing image processing system and Geographical Information System, the system modeling technique, the component development technique and dynamic link library technique. It establishes a new data model, which can process the grid data and vector data synchronously and effectively.The paper studies and analyzes the key methods of high special-differentiating remote sensing image segmentation and recognition. The software model and system model designed in this paper have used the integration technology of remote sensing techniques and Geographical Information System. It carried out the possibility of processing the grid data and vector data in one single system, which shows certain innovation in the field of geographical understanding and analysis to the remote sensing images.