Efficient automatic change detection methods and image annotation algorithm for high resolution remote sensing images

Abstract

Recent improvements in multispectral images play a significant role in analyzing the various parts of earth surface. Multispectral images are most commonly used for remote sensing applications. Land Cover and Land Use (LCLU) change relates to the changes in global proportions with unknown environmental concerns where character translation of earth s surface at local or regional scale occurs due to human activities. Land cover change on universal environmental alteration, have pervasive effects in the natural environment. Study of the land use change patterns and monitoring the changes are very important for economic planning and country. The main objective of this research is to propose an efficient automatic change detection method and semantic image annotation algorithm for high-resolution remote sensing images. In this research, proposed Maximum Likelihood Classification (MLC), efficient change detection method and modified Locally Excitatory Globally Inhibitory Oscillatory Network (LEGION) algorithm is considered to represent the urban scene by increasing the accuracy of change detection and image annotation respectively. As a consequence of urbanization in the past decade, there is increase in the global mean temperature at an unexpected rate around the world. Coimbatore, Tamil Nadu, India is one of the metropolitan cities growing towards urbanization. One of the key impacts of urbanization on environment is the effect of Urban Heat Island (UHI). Understanding the effects of LCLU pattern on UHI is crucial for reducing the effects of urbanization in cities. newlineThis study investigates the LCLU impact on UHI, based on the analysis of Land Surface Temperature (LST) in relation to Vegetation Index (VI). Furthermore, the land surface is classified into built-up, forest, agriculture, waste land and water body to know the significant changes of the vegetation area. Classification is performed with high accuracy for knowing the significant changes newline newline