Classification and change detection using Landsat TM data: When and how tocorrect atmospheric effects?

The electromagnetic radiation (EMR) signals collected by statellites in the solar spectrum are modified by scattering and absorption by gases and aero sols while traveling through the atmosphere from the Earth's surface to the sensor. When and how to correct the atmospheric effects depend un the remo te sensing and atmospheric data available, the information desired and the analytical methods used to extract the information In many applications inv olving classification and change detection, atmospheric correction is unnec essary as long as the training data and the data to be classified nl-e in t he same relative scale. In other circumstances, corrections al-e mandatory to put multitemporal data on the same radio-metric scale in order to monito r terrestrial surfaces over time. A multitemporal dataset consisting of sev en Land-sat 5 Thematic Mapper (TM) images from 1988 to 1996 of the Pearl Ri ver Delta, Guangdong Province, China was used to compare sewn absolute and one relative atmospheric correction algorithms with uncorrected raw data. B ased on classification and change detection results all corrections improve d the data analysis. The best overall results are achieved using a new meth od which adds the effect of Rayleigh scattering to conventional dark object subtraction. Though this method may not lead to accurate surface reflectan ce, it best minimizes the difference in reflectances within a land cover cl ass through time as measured with the Jeffries-Matusita distance. Contrary to expectations, the more complicated algorithms do not necessarily lend to improved performance of classification and change detection. Simple dark o bject subtraction, with or without the Rayleigh atmosphere correction, or r elative atmospheric correction are recommended for classification, and chan ge detection applications. (C) Elsevier Science Inc., 2001. All Rights Rese rved.