SPATIAL DATA MINING FOR PIXEL LEVEL CHANGE DETECTION

Chinu Jose

Abstract:Spatial data mining is a highly demanding field because huge amount of spatial data have been collected by various applications such as remote sensing, GIS, environment assessment and planning and so on. Developments of raster data capture technologies and demands from application fields call for advanced raster data analysis methods. Raster datasets usually have huge amount of pixels which challenges the efficiency of spatial data mining algorithms. In this research paper an example of land use land cover in Alwar district of Rajasthan was presented to describe the course of spatial data mining. The paper mainly focuses on pixel level identification of change that has occurred in mutlidate raster data.7 years land use land cover data of Alwar was interpreted in a single go. All the land use land cover classes were analyzed individually and the change in pixel if any was recorded. The data mining process was made much simpler by automating the task using ArcObjects. The work done in this paper can be used to enhance the ability of interpreting data to generate useful knowledge regarding the land use land cover pattern over a period of time. About the Author: Ms Chinu Jose,MTech GIS Currently pursuing MTech GIS at NIIT University BTech in Computer Science and Engineering from Mahatma Gandhi University,Kerala. E mail ID: chinu.jose@st.niituniversity.in Contact No: +91 8829844711 +91 9400454860 Recent Photograph 14 Esri India User Conference 2013 Page 2 of 5 Introduction In recent years, with the development of remote sensing and data storage technique, a great number of spatial data are generated every day, much of which is remote sensing image data. However, the use efficiency of the huge quantities of remote sensing image data is still low. It is very difficult for people to process with thousands of image data and find out knowledge from them. As the researches of data mining, information retrieval, multimedia database and other correlative field have rapid progress, it is become possible to manage and analyze large amounts of remote sensing images and find out useful information in different applications. In this work a pixel level data search has been done. Identification of change in land use patterns was the core area in this research work. The land use land cover classified image for year 2004_05 (lulc04_05) is first analyzed for a particular category of land use or land cover. Now we search for the same pixel position in all the other images. The resulting information is recorded to form the final result.

Computer Science,Geography,Environmental Science