GIS4035 - Remote Sensing and Photo Interpretation - Module 10

This supervised land classification map was made mostly in Erdas. It began as a false color image of Germantown, MD. Then the land classification was diminished to 8 different categories. The .img file was then exported to ArcMap, where the finishing touches were applied.

(The color band used was R: 6, G: 4, B: 5, as shown in the bottom right corner.)

GIS4035 - Remote Sensing and Photo Interpretation - Module 9

This map was made in ArcGis. It is an aerial photograph of the Pensacola campus of the University of West Florida. The original aerial was in true color, but was converted to the image above in ERDAS Imagine. After dividing all areas on the map into five separate categories, I calculated their areas in ArcGis, and calculated the percent coverage of each of them manually. 

GIS4035 - Remote Sensing and Photo Interpretation - Module 8

I began this exercise by contemplating the image that would be used for the final section of the project. I decided to use the previously created ETMcomposite image, due to its easy to see variation in land. I right-clicked on the layer in the table of contents, and selected “Properties.” There, I clicked on the Stretched tab and projected the image using a green-red spectrum. This gave me a clear vision of the most energy-radiating features on the map by highlighting them in red, where the deep green color would represent vegetation and low-energy areas.

 I took notice of a red area to the south-west of the main downtown area (at the center of the above image.) Judging by the amount of radiance projected by this area, it seems to be man-made. The surrounding deep-green area could be a hill-side. I went back into the properties and decided to view the image through the near infrared lens by setting the color bands to Red: 4, Green: 2 and Blue: 6. This projection told me pretty much the same thing; the surrounding red area would be natural vegetation, and area to the south would be farm fields. The feature in question, however remains the same high-radiance color of the downtown (blue.)

Once more I went back into properties and selected the following combo: red – 6, green – 4, blue – 7. This made the vegetation appear bright green, and heat bumps in red. It did not really tell me anything new. I decided (for the purpose of aesthetics) to select the Stretched tab and go with the initial spectrum-style color scheme. I selected layer 3 instead of 1, and the contrast between red and green became sharper.

To finalize the project, I went into map view and added the essential map elements. I also added a description and location of my feature, and created a grid. I exported my image as a .jpg, and exited arcmap.

GIS4035 - Remote Sensing and Photo Interpretation - Module 7

GIS4035 - Remote Sensing and Photo Interpretation - Module 6

.   I tried running the clipped aerial through several different filters in ERDAS obtained following results;

a.      11x11 LowPass – image was very blurry and difficult to interpret.

b.      3x3 Edge Enhance – image became very uniformly grey, and the lines became darker and quite pronounced.

c.       5x5 High Pass – Image became very dark, and only larger features became all grey in color. Very difficult to read.

d.      3x3 Haze Reduction – Fairly readable, but the dark lines become awfully thick on the left side of the image.

e.      3x3 Sobel – Image became monochromatic, and much too dark.

2.  After many attempts, I decided to use my initial ‘sharpen1’ file from the early stage of the assignment. Put the finishing touches on the photo in ArcMap

GIS4035 - Remote Sensing and Photo Interpretation - Module 5A

The initial raster was modified with ERDAS Imagine. A new column was added to the attribute table, that allowed for the automatic calculation of area covered by each color category. The layout as well as the essential map elements were later added in ArcMaps, and the image was exported as a JPEG.

GIS4035 - Remote Sensing and Photo Interpretation - Module 4

The idea behind this weeks exercise, was to verify the accuracy of the LULC classification done for last weeks project. We used google maps to verify the data that was previously classified only through studying the given aerial. The green points fall on to features that were classified accurately, while the red fall on to features that were misinterpreted.

Out of the total 30 points, 23 of them were accurately classified, for a total of 76.666% accuracy.

GIS4035 - Remote Sensing and Photo Interpretation - Module 3

This shapefile was created over an aerial photograph, and split into various features to correspond with the LULC classification system. All the different fields describe different land use, and land cover interpretations.

11 – Residential: The majority of the land used throughout the right half of the map seemed to be made up of suburban dwellings. Medium sized houses, with front and back yards that surrounded narrow streets. They were all confined within a single feature, and highly generalized.

12 – Commercial and Services: The more of the larger, square buildings with parking lots were labeled under Commercial and Services. Whey were mainly done so, due to their size, and proximity to a major highway.

13 – Industrial: The main feature that was reminiscent of a factory was located in the bottom right. The other two were classified so due to their size and shape.

14 – Transportation: Only one major landmark fell into this category, and that is the large highway that cuts through the aerial north-south.

43 – Deciduous Forest Land: There was a lot of forested patches of land located among the residential landmass. I assumed them to be mostly deciduous by the apparent “fluffiness” of the trees.

51 – Streams and Canals: The entire body of water on the left-hand side of the photograph was classified under this category. It was quite difficult to tell exactly what this feature was, and braking it down into more specific categories would prove very time consuming. Therefore, the more general category of “Streams and Canals” seemed appropriate.

52 – Lakes: There are three isolated bodies of water which are completely surrounded by land in the photograph. I grouped them all under lakes, as they are land locked and seemingly large enough.

54 – Bays and Estuaries: This category was given to three water features, which seem to flow into the major body of water. They are all short, and not likely to be rivers.

61 – Forested Land: This classification included every one of the overgrown islands located on the large body of water to the left hand side. Because the islands were clearly not used for anthropogenic purposes. It was difficult to decipher the exact vegetation.

GIS4035 - Remote Sensing and Photo Interpretation - Module 2

Of the two maps that we were to submit for this weeks assignment, the first one focused on defining tone and texture. We created polygons around features of our own choice, based on how dark they appeared, as well as how grainy they appeared, and organized each of them into five levels. The objective of the lab was to show understanding of what tone and texture are, and to be able to express it clearly with ArcMap.





The second part of the assignment focused on identifying features based on four distinct criteria: shape/size, shadow, pattern and association. Many of the features on this aerial could have been identified using more than one criteria, but for the purpose of this assignment we were not supposed to re-use the same features to fill the quote (3 features per each criteria.)