Kefauver_266_presentation

Kefauver_266_presentation - Comparison of EVI and NDVI for...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Comparison of EVI and NDVI for Change Detection in Papua New Guinea Presentation for ESM 266 May 21, 2005 John Kefauver New Guinea Goals See what time of year would be best for annual change detection? Would NDVI or EVI be more effective for looking at changes? Is seasonal change important? Is elevation more important? EVI vs. NDVI The Math EVI Takes dense canopies into account better.. (rho) is `apparent' (top-of-theatmosphere) or `surface' directional reflectance L is a canopy background adjustment term C1 and C2 weigh the use of the blue channel in aerosol correction of the red channel (Huete and Liu, 1996). (Thanks ATBD MOD13) Papua New Guinea EVI vs. NDVI Density Slice Information Default Ranking Red Green Blue Yellow Cyan Magenta Maroon Sea Green EVI vs. NDVI Percent Change Detection of EVI Images with Base Image 306 (Nov) and Image 001 (Jan) Blue = Negative Brown = Positive Statistics EVI 306 as initial for change detection percent 6 = NOCHANGE Day Month Year Mean stdev 1 Missing 61 Jan Feb Mar 2004 2004 2004 NA 6.05751 0.609055 NA 5.993572 0.613089 92 122 153 183 214 245 275 Apr May June July Aug Sept Oct 2004 2004 2004 2004 2004 2004 2004 6.00951 6.059413 6.086912 6.081974 6.087802 6.068885 6.015731 0.611174 0.609139 0.604963 0.606128 0.605882 0.60739 0.610539 306 336 Missing 32 60 Nov Dec Jan Feb Mar 2004 2004 2005 2005 2005 BASELINE 6.008842 NA 5.99909 6.017955 0.610794 NA 0.61208 0.611856 Percent Change Detection of NDVI Images with Base Image 306 (Nov) and Image 001 (Jan) Blue = Negative Brown = Positive Statistics NDVI 306 as initial for change detection percent 6 = NOCHANGE Day 1 Missing Month Jan Feb Year 2004 2004 Mean 6.115941 NA stdev 0.595232 NA 61 92 122 153 183 214 245 275 Mar Apr May June July Aug Sept Oct 2004 2004 2004 2004 2004 2004 2004 2004 6.02642 5.906919 5.984372 5.993439 6.048341 6.02198 6.062718 5.931994 0.606919 0.599818 0.606922 0.606709 0.605403 0.607237 0.60358 0.602919 306 336 Missing 32 60 Nov Dec Jan Feb Mar 2004 2004 2005 2005 2005 BASELINE 5.978097 NA 6.053385 6.010592 NA 0.604368 0.607129 0.60525 EVI vs. NDVI Change Detection Base Image 306 compared to Image 001 EVI Change NDVI Change The Change Detection Difference Summary Stats 306 as initial for change detection percent 6 = NOCHANGE Day Month Year Mean EVI Mean NDVI 1 Jan 2004 6.05751 6.115941 Missing Feb 2004 61 Mar 2004 5.993572 6.02642 92 Apr 2004 6.00951 5.906919 122 May 2004 6.059413 5.984372 153 June 2004 6.086912 5.993439 183 July 2004 6.081974 6.048341 214 Aug 2004 6.087802 6.02198 245 Sept 2004 6.068885 6.062718 275 Oct 2004 6.015731 5.931994 306 Nov 2004 336 Dec 2004 6.008842 5.978097 Missing Jan 2005 32 Feb 2005 5.99909 6.053385 60 Mar 2005 6.017955 6.010592 Difference (EVI NDVI) -0.05843 -0.03285 0.102591 0.075041 0.093473 0.033633 0.065822 0.006167 0.083737 0.030745 -0.05429 0.007363 Elevation Data Elevation Conclusions Vegetation indices difficult to use for large scale change detection any time of year in PNG. Elevation appears to explain patterns better, perhaps a subset for change detection would work. vs. EVI? NDVI THE END ...
View Full Document

This note was uploaded on 08/06/2008 for the course ESM 266 taught by Professor Dozier during the Spring '08 term at UCSB.

Ask a homework question - tutors are online