radiometriccorrection,partii-virginiaview.docVIP

Exercise #3 Geography 475 Digital Image Processing Radiometric Correction Part II Due: Feb. 19, 2010 In this exercise you will explore two forms of relative atmospheric correction: Improved Dark Object Subtraction (IDOS) and COST. The image and spreadsheets you need for this exercise are on the Geography server (Z:\ in GARSL). Before beginning the exercise, download the Lab 3 data onto a local hard drive. Part I: IDOS Start ERDAS Imagine and load Houston_TMimg. This is a Landsat TM scene of the Houston area (WGS Path 25/Row 39) acquired on Nov. 26, 2009. The first step is to determine a starting haze value (SHV) upon which the atmospheric correction will be based. This can be done by: 1) taking the lowest DN visible in one of the visible band histograms; 2) finding an average DN in one of the visible bands over a dark object (e.g., a deep, clear water body); or 3) visibly searching for the lowest DN actually in a scene using a pseudo-color technique. In this lab, you will use the pseudo-color technique. Open a second Viewer. Open Band 2 of Houston_TMimg in pseudo-color by clicking on the Raster Options and changing the settings to pseudo-color and Band 2. Fit the image to the Viewer by selecting View/Scale/Image to Window. Next, from the Viewer menu, select Raster/Attributes. Starting with DNs of 1, change the attribute color to yellow (do this in the attribute table by right-clicking on the black color patch and selecting yellow). In the Viewer you will see that pixels with a DN of 1 are not actually in scene. Rather, they form a “halo” around the scene’s perimeter. Continue to change image DNs from black to yellow until you find the lowest in-scene DN. What is the minimum in-scene Band 2 DN (i.e., the SHV)? Start Microsoft Excel and load the file chavez.xls. This is a spreadsheet used to calculate atmospheric correction factors generated by the IDOS model. The IDOS model is discussed in the article by Chavez (1988) and s