Image and Graphics Processing
of Computer-Aided Cartographic using Remote Sensing data
Wang Weimin,Cui
Weihong Institute of Remote Sensing Application Chinese Academy of Sciences P.O. Box 775, Beijing 100101,P.R. China Abstract: To realize automatic extraction and mapping of remote sensing information, it’s an important technical means that adapt the development of remote sensing technology system and speed up the remote sensing application . This paper introduces how to use the technical method of computer classification and automatic mapping to make the thematic map of land use; then describes the test and choice of the composite about different types of remote sensing data and the method of classifying types, the constructing method of polygons boundary; finally, discusses the features of the methods . The methods are achieved on the Microcomputer Image and Graphics Processing System . The research on remote sensing, cartography and Geographic Information System (GIS) we are engaged in is right within the category of information sciences. Remote sensing is a source of global information, map is a carrier of spatial information and GIS relates them to each the to form information flow and have the capacity for storing, retrieving, analyzing and quick displaying the information flow and supporting decision making. Therefore, we should strengthen the connection and coordination among them and shape an overall iaea of the integrated discipline system to enable this spatial science to benefit our economy and society more. The emphasis of this paper is to discuss several methods of image and graphics processing in Remote Sensing thematic cartography . Image processing and graphics are two important steps in R.S information automatic cartography . The former, by which thematic information is obtained, is the basis for quality of map compilation, while the latter is, according to determined objective and fixed rules, form new information combination by careful sieving, technology and analytical processing. Obviously, there is a little difference between them both on definition and on denotation. However, with the emergence of digital image and digital map, it’s possible, by the medium of computer and digital techniques, to realize conversion and mutual-supplement of image and graphics data. So, from our point of view , the critical problem in Remote Sensing automatic cartography is the data processing techniques, in which the conversion between raster and vector data, patterns is the key problem. Expert System method and Intelligent Theory is very useful for the solution of data conversion. 1. Study of image processing methods . The application’s potentiality of Remote Sensing images depend not only on the properties and qualities of images but also on the methods used in image processing. Because the classification and extraction of R.S image thematic information has a close relation to the distinction of object boundary, contour enhancement plays a great role in image processing. There are several methods that we can use to carry out contour enhancement processing, but the results of these methods are not satisfactory, because the noises also enhance with processing. Here , we improve the parameter high pass filter and obtained satisfactory results. The equation is as following : G(I,j)=P(I,j)+f(Dp)+(A-P(I,j)(1-B) Where: P(I,j) is the grey value of the central pixel in the chosen window; F(Dp) is function of contour enhancement; A is additional offset to high pass filter; B is a constant, ranging from 0.0-0.1; Dp=P(I,j)-M(I,j); M(I,j) is the mean grey value of the window in which (I,j) is the central pixel. The value of Dp changes with the boundary areas than in inside areas; Function F(dp) can be determined according as the boundary as sharp, medium-sharp or smooth. So, image processing can be optimized by several ways of contour enhancement. 2. Processing and Mapping of R.S Image Information Although R.S image actural reflection and reproduction of spectral characteristics of ground objects, it’s difficult to use then directly in cartography, and so, in addition to traditional processing techniques, it’s necessary for use to make a set of cartographic processing aimed at image production , such as modification of mathematical base, sub-image processing. Satisfactory result will be obtained by using expert system theory. 3. A new thinking used in vectorization of raster data Differing from traditional ways based on through-searching and tracing conversion, our method is by computing coordinate chain of data in which artificial intelligence and logical thinking have been used so that is ensures us to find the best choice of efficiency and quality of vectorization. The method includes three steps:
The countours extracted like that is exact, but because they combined some parameter both of raster data and of vector data, it is necessary to remove the trace of raster data, such as non-sequence and non-continuity in order to make vector mapping . We can process the tracing and node-matching by following steps: (1) Read a set of data from the boundary coordinates data file and record Its characteristics code. (2) According to eight kinds of patterns and six classes of combination and starting from a actular coordinate points, begin searching in certain range around this point. (3) Once the point is searched, match this point with node coordinate and record it in a new file, meanwhile set its characteristics code to be zero and start searching again from this point. (4) Repeat (2) and (3), when the polygon closed, switch to another group. (5) Repeat from (1)-(4) until the characteristics code of all the data have become zero. So vectorization is finished. 4. Map compilation. By means of above conversion and processing, thematic image data will be converted to purified, classically united vector mapping data which is composed of characteristics code and couples of coordinates. These data can be stored in bank directly, they also can be displayed, copied or mapped, but it’s necessary to make further compiling or processing which mainly includes following work:
Conclusions . Based on these description above, one method like building blocks structure is presented in this paper which will generate information maps of remote sensing automatically. Organic links between image and graphics processing and automatic cartograph have been implemented. Application of this method have been included in the task of land-use dynamic monitoring and urbanization spatial analysis in Beijing city and present land-use situation mapping in loess plateau. All the applications indicated that the above method not only provide a effective means to use space remote sensing as a important mapping information source, but also has very important significance in the improving of patter cogization and thematic analysis, and in the providing of resource list, dynamic data and thematic maps timely. Reference:
|