A Study Of The Technology Of Areal Ground Features Change Detection Based On Vector Map Data

To meet the requirement of the currency of geospatial information, how to realize fast and automatic change detection for geospatial data becomes a problem to be urgently solved. The change detection for areal ground features is an important part. In this paper, the change detection technology for areal ground features based on vector map data is studied, and the main subjects investigated are as follows:1. Analyzing the present status, difficult problems and research direction of feature on the change detection and the extraction technology applied in change detection. Putting forward the change detection flow of areal ground features based on vector map data.2. Studying the method and model of image segmentation. Putting forward an image segmentation method which adaptively determines the segmentation threshold value and applies mathematical morphology to process the images segmented. The image segmentation model is immunogenetic algorithm based on max.entropy and Otsu.3.Studying semi-automatic extraction of areal ground features. Putting forward the habitation textural extraction based on Gaussian ambiguity, the water area adaptive extraction based on grey level and vegetation adaptive extraction based on equalizing value of grey levels. And comparing the results with the old extraction methods.4. Studying raster data vectorization and vector data rasterization. Putting forward vectorization method based on genetic algorithm and compressing the data using Douglas-Peucker vector data compression algorithm. Putting forward a vector data rasterization method which is a boundary monument scanning beam rasterization method which combines the boundary monument with scanning line seed.5. Studying areal ground features change statistic method based on vector map data. Putting forward the change statistic method based on acreage, and counting the varieties of vector map data by using polygon intersection of vector data and changed pixel of raster data. The experiment shows that the two statistical methods are efficient.6. Building the evaluation model based on areal ground features change detection on vector map data, and putting forward a method to evaluate the result of change detection by using the theory of fuzzy comprehensive judgement.