Forest fires are common feature in Andhra Pradesh forests, causing adverse effects on eco systems, environment and humans. Near real time forest fire detection in India has been carried out by National Remote Sensing Centre (NRSC), Hyderabad and Forest Survey of India (FSI), Dehradun using thermal bands of MODIS TERRA and AQUA satellite datasets. MODIS is a sensor on board TERRA and AQUA satellites and has 36 bands with resolutions 250 m, 500 m and 1 km. In this study, forest fires can be detected from MODIS TERRA Land Surface Temperature (LST) product. The MODIS Terra Land Surface Temperature and Emissivity product (MOD11A1) provide daily per-pixel temperature and emissivity values in I km spatial resolution. The r.modis suite has been used to import MODIS satellite data in GRASS GIS. The r.modis uses the pyModis library and the MODIS Reprojection Tool software to convert, mosaik and process MODIS data. K-means clustering has been used to segment the image into identical regions based on wavelength and measure to locate the high temperature region of the study area. Wavelet transform has been used to compute the mean wave length, plays an important role to detect forest fires and this process will be carry out using python based modules. Accuracy will be computed based on the MODIS active fire hotspots of the study area.
Copyright © 2015. All Rights Reserved.