AUTOMATIC LAND COVER CHANGE DETECTION BASED ON IMAGE ANALYSIS AND QUANTITATIVE METHODS

摘要

The methods of Automatic land use and land cover change detection based on remote sensing image has been widely applied in research for LUCC (Land Use and Cover Change), nature resource management and environment monitoring & protection. Taking into account that multi-temporal remote sensing data also were used to retrieve important information for change detection, and several relevant algorithms have been developed, e.g., Image Differencing, Principal Component Analysis, Post-classification Comparison and Change Vector Analysis etc. to automatically detect the changed information. However, while changes are automatically monitored, the case that time one (T1) data is existed land use and land cover maps, and another time (T2) data is remotely sensed imagery is also very common. Under the condition that one time (T1) data is existed land use and land cover map, and another time (T2) data is remotely sensed imagery, how to detect change automatically is still an unresolved issue. The frequently adopted method is to interpret the registered and superimposed T1 data and T2 data, which is time consuming and has a big workload. In fact, the unchanged information is the main information. Therefore Tl data has a great deal of land use and land cover information consistent with T2 data. If the useful information is mined by the computer, and the knowledge database of land use and land cover classes based on the statistic information is established, the changed information can be automatically and quantitatively detected via the guide of the knowledge. This paper developed a land use and land cover class knowledge-oriented method for automatic change detection under this situation. Firstly, the land use and land cover map in Tl and remote sensing images in T2 were registered and superimposed precisely. Secondly, the remote sensing knowledge database of all land use and land cover classes was constructed based on the unchanged parcels in Tl map. We can make good use of many different methods of image analysis to extract information of texture, conformation, color and so on. At the same time, we utilize the methods of quantitative Remote Sensing to compute some important parameters, such as LAI (leaf area index), NDVI, reflectivity, soil water content etc. we can input all these information into knowledge database according to land cover classes and extracting method. Thirdly, guided by T1 land use and land cover map, feature statistics for each parcel or pixel in remote sensing images were extracted. Finally, land use and land cover changes were found and the change class was recognized through the automatic matching between the knowledge database of remote sensing information of land use and land cover classes with the extracted statistical information in that parcel or pixel. Figure 1 shows the work flow of automatic land cover change detection. Experimental results and some actual applications show the efficiency of this method.