Lab 4: Enhancing imagery (Due by 4pm Feb 15)

Introduction

In this lab you will learn how to enhance satellite imagery using spatial, spectral, and radiometric techniques.

Instructions

Read pages 155-181 of the ERDAS Field Guide (FieldGuide9_1.pdf). You can skim the topic of spatial convolution for now. It will be covered in Week 7’s lecture. Open the Tour Guide (TourGuide9_1.pdf) and skim the Image Interpreter section (pages 257-282). For this lab, you will have to copy the following files to your workspace in the C:\users folder. You can find these files in I:\Students\Instructors\Geoffrey_Duh\GEOG4581\Lab4.

dmtm.img

lanier.img

loplakebedsig357.img

panAtlanta.img

tmAtlanta.img

Complete the tutorial exercise in the Image Interpreter section (pages 257-282, stop at Wavelet Resolution Merge) and produce pictures of the following images - convolve, crisp, reverse, inverse, tasseled, and mineral. Make sure you clearly annotate the pictures and submit them with your report.  Answer the questions at the end of this exercise.

Additional Notes

1.      When doing “subsetting an image without snapping”, you won’t be able to keep the coordinates you typed in. However, the more you zoom in on the image, the less displacement of the coordinates you will get. Make sure you zoom in enough to see individual pixels, so that you can see the effects of subsetting an image with and without snapping.

Questions

  1. What is "snapping to raster" when subsetting an image? Make two figures showing the effects of subsetting with and without snapping. The figures must show blowup views of one corner of the inquire box superimposed on the subset and original images.
  2. Click on the AOI tool within the subset dialogue and explain the three different options and the different results you get using each.
  3. Fill the table with the coefficients of a 3 x 3 Edge Detect filter.

 

 

 

 

 

 

 

 

 
  1. What are the layers of a tasseled TM image? Describe what several features (e.g., water, urban, forested, and agricultural) look like in the false-color tasseled image and why they look that way.