Spectral Analysis and Identification of Vegetation Types | GEOG 4980, Lab Reports of Geography

Download Spectral Analysis and Identification of Vegetation Types | GEOG 4980 and more Lab Reports Geography in PDF only on Docsity! Lab #8: Spectral Analysis and Identification of Vegetation Types (Due on 4/6/2004) Objectives  To find an effective method for labeling vegetation covers in a remote sensing image. Assignment You are given the following Landsat 5 Thematic Mapper image, imperialvalleyTM.img, to work with for this lab assignment. The image was acquired on December 10, 1982 and covers a portion of the Imperial Valley in California. The imagery represents four vegetation types: sugar beets, alfalfa, cotton, and fallow. The actual image data can be retrieved from S:\GEOG\ FACSTAFF\jminhe\RsData. Also your textbook provides a fuller description of this image and phonological cycles of different crop types in the study area, especially figure 10-20 (Page 360), plate 10-4, and plate 10-5. Review these sections and illustrations before attempting this assignment. RGB = 321 RGB = 432 RGB = 532 RGB= 732 GEOG 4980/5960 - Introduction to Remote Sensing 1. View the image using ERDAS Imagine and different viewing methods (hint: multiple viewers, different band combinations) 2. Notice that the image has an orientation different from the pictures in this handout and those in your textbook. Suppose the orientation of the ground reference map on page 360 is correct, what might explain this difference? (Hint: Read more about the U.S. Landsat 5 satellite in your textbook.) Where in Imagine (and in the textbook) can you find whether this image might or might not have a correct orientation? 3. Create a graph using Landsat TM's seven bands (Band 6 and band 7 have been switched in the current dataset) depicting the spectral profile for a) sugar beets, b) alfalfa, c) cotton, and d) fallow fields. After careful analysis of each of the spectral curves, choose the band(s) that provide the best discrimination among these four crop fields. 4. Jensen in his book demonstrates how to derive a classification map of the crop fields in the image through a manual approach. He also mentions a computer-based classification. By comparing the classification map (the second figure in this handout as well as Plate 10-5e in the textbook) to the ground reference map (on page 360), we may find some obvious errors! Identify and label the misclassified crop fields and provide some explanation why this could be wrong. (Hint: computer classification is based on the spectral differences among crop types. This difference must be large enough for the computer to separate crop types. How can you examine and demonstrate that a misclassified crop field (or a pixel in the field) has a larger spectral difference from its true kind and a higher spectral similarity to the crop type it was misclassified to?) Suggestions and requirements: 1. You should employ the Imagine tools you have learned so far to explore the image and look for information relevant to the questions. 2. You are encouraged to explore and use new tools. 3. Your explanation can be very much enriched if also based on information from the textbook. GEOG 4980/5960 - Introduction to Remote Sensing