A Modified Landscape Expansion Index Algorithm for Urban Growth Classification Using Satellite Remote Sensing Image

Traditional landscape indices are only able to specify the characteristics of urban area at one particular time. The same landscape index applied at different time would generate different result. A new landscape index, called landscape expansion index (LEI) is able to reflect the specific growth type of the urban area using multi temporal datasets. This paper evaluates the efficiency of LEI algorithm for urban growth classification using several Landsat Thematic Mapper images of Klang Valley, one of the most rapid urban growth areas in Malaysia. The images are pre-processed into binary images containing the undeveloped and existing region of the study area. The identified new urban region are then classified by incorporating LEI in the classification algorithm. Result shows that LEI has the limitations to properly identify infill and expansion growth. A modified LEI is proposed by taking into accounts two parameters: location of the new urban region and the existence of existing region surrounding it. Experiments with all available images produced positive output whereby the modified LEI is capable of correctly classifying urban growth types.