GEOG 430

Remote Sensing

Syllabus

Required readings

Lecture Information

• Week 1: history, d'oh... missed a day due to snow...
• Week 2: lecture on resolution, And a closer look at landsat data, including how to get data off earthexplorer. reflectance, data, and stretching. Sensors.
  • End of Drop Add on the 10th.
• Week 3: atmospheric corrections.
  • No class monday (MLK)
• Week 4:
  • Filters and resampling. Watch these three spatial filters videos (first vid). Also (another lecture from the Netherlands. worth flipping through if only for the examples).
  • Georectification.
• Week 5. Digital image processing. Looked at Earth as Art book.
  • 2 feb - end of 50% refund withdraw from course.
• Ratios (including looking at a lot of different imagery using ratios)
• Reflectance curves
• Week 6:
  • Midterm exam on Monday, 7 February!
  • Principal Component Analysis.
  • Drones
• Week 7: Unsupervised Classification, supervised classification.
  • movie on wed (and a little of friday)
  • 18 Feb - end of uncontested withdraw from course
• Week 8: accuracy assessment.
  • No class monday - president's day.
• Week 9: LIDAR and moretime to work on your projects
• Week 10. Radar
• Final exam. Wednesday, 16 March, 8-10AM

Tutorials: Note, labs handed out on monday will be due friday, Wednesday labs will be due the next monday. And friday labs will be due the next Wednesday!

• Lab 1 - We'll start with a little review from airphoto. Work through the "Eye in the Sky" book (available in the lab) and answer bonus question 1 from each "chapter". 3 points. Please don't write in the books - they are collector's items and can no longer be purchased. Must be typed!
• Lab 2: The lab and the worksheet 3 points. (obviously, fill out and turn in the worksheet. Hardcopy, my office door). (since there's no class the next monday, this will be due tuesday).
• Lab 3: the lab and the worksheet 3 points
• Lab4 - in class exercise looking at different sensors. Due on friday of week 3 (not the next monday). 2 points.
• Lab 5: Pan Sharpening. 2 points.
• Lab 6: Georectification. 3 points.
• Lab 7: Filters and ratios. 3 points.
• lab 8: PCA. 3 points.
• Lab 9: Unsupervised classification. 3 points.
• Lab 10: supervised classification. Part one. 5 points.
• Lab 11: Accuracy assessment of your supervised classification. 4 points.
• Lab 12: LIDAR. Yup, frickin' planes with frickin' laser beams. 3 points.

Project: The overall goal is to track landcover/use change in an area over, say, 20-30 years

• Step 1 - think of an area that has changed over the past few decades or so. The world is your oyster. Think at least 20 miles by 20 miles - but must exclude WA State. Write up about a paragraph which describes your study area/topic (what you're mapping and where). due by 11 February. 1 point
• Step 2 - get data! Simplest would be landsat, but feel free to use any sort of applicable imagery. The first image must be from the last year or two; the second image must be AT LEAST 20 years older than the current image. Head off to http://earthexplorer.usgs.gov/ - this is the best (but not only) place to get all sorts of imagery. If you want landsat, be sure to search the Landsat collection 1 level 1 section. If you wish to be more creative and find different sorts of imagery from different places, go for it. Remember, the two images MUST be taken at the same time of year.
• Step 3 - get the data into ERDAS. Note, the erdas importer should work on the xxx.tar.gz file you downloaded. Be patient, the importer takes time. When in the importer window, make sure you import both the PAN image and the MSI image - you don't want the thermal bands, trust me.

  A few of you are having problems importing older landsat data......  After a google search, turns out erdas's importer is still buggy (you would think they would fix this...).  So, a workaround.

  • First, make sure you are downloading the right landsat data.  In earthexplorer, search the Landsat collection 1 level 1 section.  And select Level-1 GeoTIFF Data Product  when you get the download option.
  • OK, now unzip that file using 7-zip - you should have a bunch of .tif files and some metadata.
  • Onto ERDAS - goto raster - spectral - layer stack.
  • Open the input file dialog, change file type to tif
  • Select band one
  • OK
  • Add
  • Select band two
  • OK
  • Add
  • Repeat until all bands are in.  Make sure you do it in order!  Hit OK and it *should* work. 

Make sure the two images line up properly in your study area. Again, DO NOT load/use the thermal imagery into your band stack (if using importer, you will want the msi image, not the msitir image).

On 21 February, I will check to see that you have both images downloaded and ready to use in ERDAS- so make sure you bring everything to class that day. Worth 2 points.

• Step 4, clip out your study area. Load both images (old and new). Goto raster, subset and chip, create subset image. input your input and output files. highlight the "snap pixel edges to raster image" Then build an inquire box: On the Home tab, click Inquire > Inquire Box to open an Inquire Box in the active View. Resize the box to fit the area in which you want to work. Now, back to subset.... click the "from inquire box" button. You should see the map coordinates change to fit your inquire box. Finally, make sure the input and output data types are the same and that you have selected the right number of layers. Do it. Then repeat for the second image (keep the inquire box up! Make 100% sure that you clip out the identical areas).
• Step 5 - Figure out which landuse/landcover classes exist during both dates (note, there may be ones that are in one image that aren't in the other, if enough change has happened). Follow your book, page 611 - 618. Or goto https://pubs.usgs.gov/pp/0964/report.pdf and take a look at table 2 on page 8. You should be able to get to Level II.
• Step 6 - do a supervised classification of your images (we'll get to this later in the quarter). Make sure your classes have the same value (number) in both images. This will be important when you compare the two dates. Don't forget your accuracy assessments - they should look just like those from the recent lab.. (Hand in a quick writeup which includes both the original and supervised images and both confusion matrices (formatted like you should have for the accuracy assessment lab). Yes, I expect readable color images with appropriate cartography (neatlines, N arrow, scale, etc). Note, if you prefer, you can always do your cartography within Arc.
• Step 7 - Carefully examine both images and describe the change you see. Is it real or an artifact of classification? Go through this for every class. In other words, tell me how and where each class changed over the past xx years (for example, the old growth forest class from 2002 changed to the young forest class in 2022, primarily in the SW part of the study area. Or - we saw about a 50% increase in the urban area, as forest and range lands to the north of <city> were developed.)
• Give me a final paper which includes
  • An introduction which includes your question (what you're doing and why),
  • Descriptions of your data - year, sensor, time of year, where you got it, etc.
  • your methods (what steps you took),
  • your results,
  • and a final "thoughts" section which details what worked, what didn't, what you would do differently next time, etc.
  • Note, your final writeup should include a lot of images, and they all need to be readable (no postage stamp images - fill the page!). It's better to give multiples than too few.

Due Dates

• Step 1: 11 February. 1 point
• Step 3: 23 February. 2 points (checked in lab that day on computers)
• Step 6. 4 March. 2 points
• Final writeup: Due the friday before finals week: 11 March. 10 points

Some potentially useful links for you:

• Remote sensing core curriculum
• NASA images and animations
• Gorgeous pics: Earth as Art book.