High Hot Spots Recognition Technology Detecting Indigenous Coke-Making Sites Using Remote Sensing Data
|School||Chengdu University of Technology|
|Course||Earth Exploration and Information Technology|
|Keywords||active fire detecting indigenous coke-making sites hot spot detecting near-infrared wavelength thermal-infrared wavelength sub-pixel unmixing non-isothremal pixel land surface temperature retrieval|
High hot spots distribute anywhere either on the earth surface or in the space such as forest fire, grassland fare, active volcano, coal fare. These hot spots cause severe environmental and economic problems. Although satellite remote sensing has been used successfully to detect these spots, a satellite data based concept that can quantify the majority of detected hot spots is still missing. There are many problems about hot spots detecting mechanism using satellite data. This dissertation tests a new generation of satellite instrument TM sensor which explores the potential to determine hot spot radiative energy. This dissertation made choice of Shanxi province as research area. An attempt to detect and monitor indigenous coke-making sites was achieved based on Landsat TM data. The key conclusions are as following:(1) High hot spots recognition is affected by the solar radiance, spatial character, spectral character, radiance transportation, aerial effect and sensor respond. Analysis procedure of high hot spots is presented for improving the recognition accuracy. This method is important for next research.(2) Put forward an method for high hot spots recognition using the near infrared (0.7-3.0μm) remote sensing data. The theoretical analysis of Landsat TM sensors outlines the fact the thermal infrared or near infrared (0.7-3.0μm) are effective in registering high temperature spots. TM7 and TM5 have advantage to detect very high temperature spots through some researrch cases wildfire, coal fare and volcano eruption were presented in dissertation.(3) Indigenous coke-making sites cause severe environmental and economic problems. This dissertation first explores the potential to detect coke-making site based on Landsat TM data (day time). Although TM7 and TM5 spectral radiance of an image contains not only hot spot temperature but also solar reflectance; they are sensitive to be able to register spectral indigenous coke-making radiance. The thermal infrared (TM6) is, however, particularly effective in registering these hot spots. In this dissertation, a new method is presented, which links the indigenous come-making radiance to recognition through TM751 band combination. During five years (from 1999 to 2004 at Shanxi Province), the quantity of indigenous coke-making sites distributed mainly on the plain increased by 10 times and the figures given by local authorities can in fact be approximated. It is thus expected that Landsat TM data will become a crucial tool in obtaining reliable, quantitative information for indigenous coke-making sites.(4) The present dissertation presents an improved method for surface temperature retrieval algorithm of sub-pixel indigenous coke-making hot spots. The thermal data saturate for temperatures above 68℃and indigenous coke-making sites occupying less than a pixel requires modification of the methodology. The temperature estimation from the method is found to be in good with agreement with ground observations.