ABSTRACT

In the present work multi-spatial, spectral and temporal images taken from space platform were acquired to analyze Talakadu (an archaeological site in Karnataka, India) area and identify linear features and geometrical shapes on the surface that are not natural but can be human artefacts from the past and therefore might indicate potential locations for archaeological exploration. Before such remote sensing data can be used they have to be processed and prepared for facilitating meaningful interpretation. The scope and importance of image processing and enhancements depend on the surface conditions of study area. The peculiar geographical configuration of the Talakadu area has made the greater part of an old township completely buried under wind blown sand from the river Kaveri that skirts the west of Talakadu. Towards carrying out analysis in spatial, spectral and temporal domains, space imagery from Landsat Thematic Mapper, Multispectral images from LISSIV sensor of Indian Remote Sensing Satellite (IRS) series and high resolution images from IRS CARTOSAT series were used. Several image-processing algorithms were employed to evaluate their efficacy to identify linear features and geometrical shapes that might indicate presence of buried remains. These include enhancements and improvements in data quality, Pan sharpening, Principal Component Analysis for enhancement and change detection studies. These analyses were conducted after subjecting all the imagery to a common reference frame through automated registration using non-linear methods with sub-pixel accuracy. This work presents the details of the various image processing techniques used to investigate their efficacy in the context of identifying the different archaeological details both known ones and possibly new potential locations. Our preliminary analysis indicates at least four likely locations for further archaeological exploration, thus suggesting a new approach in the light of the methodology available.

KEYWORDS

image processing techniques, archaeology, Talakadu site, image registration.