My research work is on application development (i.e., Colorimetrically Improved Classification Accuracy) in the field of Super Resolution Mapping. This proposal is classification of satellite images, with the specific aim to construct classification maps at a finer resolution than the spatial resolution of the satellite image. For this, information is required at the subpixel level, contained within the multispectral/hyperspectral pixel itself or from other image sources of high spatial resolution. In this proposal, we discuss different methodologies for subpixel satellite image classification, based on the fusion with contextual information obtained from a high-spatial resolution (panchromatic or color) image. In one strategy, contextual features (Morphological Attribute Profiles) are calculated from the color image and along with spectral features from the satellite image applied in an optimised sub pixel level fusion classification by using colorimetry. In the other strategy soft classification or unmixing along with subpixel analysis and arrangement is applied on the satellite image to obtain fine resolution maps. The color image provides contextual information for the subpixel analysis.
The perception-oriented measures such as the CIE color difference metrics and the spectral measures such as the Euclidean distance between high-dimensional vectors that measures in the first category usually define particular viewing conditions (illuminant, standard observer) whereas spectral measures on the other hand, allow comparing full reflectance spectra in an unconstrained fashion. In this research proposal, we consider both perceptual and purely physical properties of the scene to detect salient objects.
The Euclidean distance between pixels (CIELAB trichromatic values). The conversion from reflectance to tri-stimuli values with respect to the CIE standard observer data, or so-called Color Matching Functions. These are indeed the standard assumptions when it comes to viewing conditions in the literature. Tri-stimuli in XYZ were then converted to the perceptually pseudo-uniform CIELAB color space. This feature includes both lightness and chromatic information.
Paper Published (National/International)
- Suresh Merugu, Kamal Jain, 2013, “Colorimetrically Resolution Enhancement Method for Satellite Imagery to Improve Land Use” 14th ESRI User Conference id: UCP0046, New Delhi, India, 11-12th Dec, 2013.
- Suresh Merugu, Kamal Jain, 2013, “To Generate High Resolution Images (Conventional Subpixel) from Low Resolution Satellite Images: Colorimetry Concept” ISG &ISRS, Visakhapatnam, India, 4-6th Dec, 2013.
- Suresh Merugu, Kamal Jain, 2013, “To Detect Infrastructural Damages Caused by Natural Disaster using Low Resolution Images” Joint International Workshop of ISPRS & INCOIS WG VIII/1 AND WG IV/4 on Geospatial Data for Disaster and Risk Reduction, Hyderabad, India, 21-22nd Nov, 2013.
- Suresh Merugu, Kamal Jain,2013 “Sub Pixel Analysis on Hypothetical Image by using Colorimetry” International Journal of Recent Technology and Engineering (IJRTE), ISSN: 2277-3878,Volume 2, Issue-4, September 2013.
- Suresh Merugu, Kamal Jain,2012 “Development of Sub- Pixel data based on Colorimetric Approach” at National Symposium on Space Technology for Food & Environmental Security at Annual Convention of ISRS & ISG, New Delhi, during December 5-7, 2012.
