Laserfiche WebLink
MEMO 89.2 <br />The digital numbers representing dark objects were obtained empirically. Image statistics were computed <br />and the resulting image histograms analyzed. The digital number selected for the dark object in each <br />band was the lowest value at the base of the slope of the histogram. If the slope of the histogram was <br />gradual (i.e. contained very few pixels), these pixels were ignored and the digital number value was <br />selected where the slope of the histogram increased significantly. This information, along with the other <br />parameters of the equation described above, was introduced to atmospheric correction models developed <br />in ERDAS Imagine specifically for each Landsat scene. An example of the radiometric and atmospheric <br />correction models developed in ERDAS Imagine is shown in Appendix A. Figure 3 provides an example <br />of results obtained from the radiometric and atmospheric correction procedure. <br />~, *... <br />{ e ~E i~2 r <br />9 A.~ Ydf ~d ,& ~ r <br />>, ~ <br />~~ ~ <br />~~ <br />~`, <br />near <br />infrared band combination (4-3-2) <br /> I <br /> ~ <br />~ _ <br />}.r ~, ~ <br />' <br />'~ , 1,,,~ ~~7e' <br />~ <br />y \ <br />a a,_.. <br />K ~~~~~~~ <br />x'4 <br />$ ~ ~ <br />~ <br />~, a~: __ <br />~ <br /> _~ <br />~~ r <br />~~ = ~. <br />t <br />m '*~ 4 <br /> <br />h x <br />E~ <br />4~ <br />(b) Radiometrically corrected Landsat TM in <br />near infrared band combination (4-3-2) <br />Figure 3. Comparison Between an Uncorrected and a Radiometrically Corrected Landsat TM Image <br />2.3 Image Enhancement <br />2.3.1 Normalized Difference Vegetation Index (NDVI) Transformations <br />As described in SPDSS Memorandum 89.1, multi-temporal NDVI images produced by the composition <br />of individual date NDVI for each Landsat frame were used to provide additional information for manual <br />photo interpretation of irrigated parcel boundaries on the DOQQs as well as in subsequent processing for <br />crop classification (for more details on photointerpretation of irrigated parcels boundaries please refer to <br />SPDSS Memorandum 90.2). <br />NDVI was derived from the difference between the near-infrared region of the electromagnetic spectrum <br />(e.g. Landsat TM Band 4), and the visible red region (e.g. Landsat TM Band 3) using the following <br />equation: <br />NDVI =NIR Band -Red Band Where <br />NIR Band + Red Band <br />NIR Band =Near infrared reflectance <br />Red Band =Red reflectance <br />The results of the NDVI operation were floating point images scaled from minus one to positive one, the <br />theoretical range of the index (minimum and maximum values for the given NDVI image). NDVI values <br />below zero indicate areas without vegetation cover, while NDVI values above zero indicate the presence <br />of vegetation. The higher the NDVI value, the greener and more vigorous the vegetation is on the <br />ground. Floating-point representations are slower and less accurate than fixed-point representations (they <br />Page 5 of 45 ~RFversfde TecAnotogy, fnc. <br />4'JaYCr Resources Errgi~ecr:np an~i CansaFlrnp <br />