Tuesday, October 17, 2017

Lab 6: Spatial Enhancement

This lab focuses on uses spatial tools to enhance an image. This lab was composed of 3 exercises. The first exercise was mainly about unzipping data and importing the file into ERDAS Imagine and converting it into the correct format so the image can be used.

The second exercise is where we learn about spatial enhancements. The first part was using the Convolution tool and setting a 3x3 Low Pass kernel on a raster. This type of kernel gave the image a blurry look while highlighting the changes over a broader area. After that, I then used a 3x3 High Pass kernel on the original image. This type of kernel does the opposite of a low pass kernel; highlighting smaller scale differences. The next part was exploring the Focal Statistics tool in ArcMap. In this tool, I used a rectangular 7x7 Mean type kernel. I also used a 3x3 Range kernel.

The third exercise was based on image enhancement. The original image had striping across the image. In ERDAS Imagine, I used Fourier Transform tool to reduce the amount of striping, and then used the Convolution tool with a 3x3 Sharpen kernel. After that, I opened this image in ArcMap and used the Focal Statistics tool to create a 3x3 MEAN type kernel to enhance the image as much as I can. Below is my final deliverable showing how much I enhanced the image.

Thursday, October 12, 2017

Mountain Top Removal: Report

This post concludes the final portion of the Mountain Top Removal project. After analyze week, I converted my classification raster into a polygon layer. I then removed "noise and interference" by deleting any ridges or water features that may have mistakenly been classified as MTR areas. This helps create a more accurate classification. I then created an Acres field to my polygon layer, and calculated the amount of acreage that was classified as MTR.
After that, I conducted an accuracy check, where I randomly generated 30 points and cross-referenced them to the original Landsat image and calculated the accuracy percentage.

I then packaged my file into a Map Package and sent it to the group leader, who then compiled all of the data together and calculated the accuracy percentage of that and how many acres of MTR there were. He then packaged that and sent it to the rest of the group.

Here is a screenshot of my final map that I created on ArcGIS online. This includes the compiled polygon layers and the stream feature created in Prepare week. I also included U.S. Coal Fields to compare the MTR areas to actual recorded areas.

Overall, I think that this project was very interesting and I think that our group did a good job analyzing and compiling all of the data. The only limitations was the cloud cover in the Landsat images, because they were mistakenly classified as MTR, and so were their shadows.

I also completed the story map journal, which basically explains the whole project, including the study area, background information, and the analysis, while also showing the completed map, which is actually interactive. My completed map journal can be found here.

Monday, October 2, 2017

Lab 5a: ERDAS Imagine

This weeks topic was about the program called ERDAS Imagine and learning about this program and how to use it. The first exercise of the lab was just calculating math about wavelengths and frequencies.

The second part of the lab was to gather an understanding of the program; learning the basic tools and adding images and opening the attribute table, etc.

The third part was about map making, and preparing the data in ERDAS first, and then completing it in Arcmap. In ERDAS, I added a classified image of the forests in Washington state, and opened the attribute table and created a new area column in this table. After that, I used the Inquire box to create a subset map of a small portion of the image. I saved this image in my output folder, and then opened it in Arcmap. I fixed the symbology and added descriptions to show the area for each classification. Below is the map I created.

Friday, September 29, 2017

Mountain Top Removal: Analyze

For this week, I learned about the ERDAS Imagine program, and applying classification methods to a raster. Each person in my group analyzed their own Landsat. With this landsat image, I used ERDAS to perform an unsupervised classification, with 50 different classes. After the process was complete, I then manually clicked on different pixels and assigned them a new color based on if they were MTR or NonMTR, leaving me with 2 classes. After that, I opened this image in ArcMap and reclassified it so only MTR polygons were visible. My screenshot below represents all of the MTR areas of my landsat image.

Tuesday, September 26, 2017

Photo Interpretation Lab4: Ground Truthing

This module's lab worked off of the map from last module. In the last lab, I classified parts of Pascagpula, Mississippi. This includes classes like residential, commercial, lakes, forests, etc. Well, this week was about "ground truthing", which basically means I determined if these classifications were accurate or not.

To do this, i first created a new point shapefile and added two fields to the attribute table; "True_YN" and "New_Code".  I randomly selected 30 areas on the map, and then manually found these areas in Google Maps and I used Streetview to actually look at different buildings in Pascagoula.It was a little challenging, but also interesting to see what buildings actually turned out to be because it's hard to tell what something is specifically just from a top-down view. 

After determining if the sample point was accurate to the original classification, I put a Y under the "True_YN" field. If it was not accurate, I put an N, and reclassified it and wrote the new code in the other field. 

Below is the map I created. It's very similar to last labs map, but this one shows the random samples and whether they were accurate or not to that classification. I also included the ground truthing accuracy.

Friday, September 22, 2017

Mountain Top Removal: Prepare Week

This module's project is based on mountain top removal on the Appalachian Mountains. For this part of the project, prepare week, I had to transform the DEM layers into a polyline shapefile representing different stream features. This was accomplished by using different Hydrology tools under the Spatial Analyst toolset. Below is the basemap I created for this project. This includes the stream features, the water basin polygon layer, and the study area boundary.

This project also consists of creating a story map tour and a story map journal. 

The story map tour presents the 6 stages of mountain top removal, and a brief description of each stage. You can view this tour here.

The story map journal is currently a work in progress, and will be completed during Report week. For now, most of the journal has the correct sections, but holds placeholder images and texts until the Analyze portion of the project is completed. You can view this journal here. 

Monday, September 18, 2017

Photo Interpretation: Land Use & Land Cover

This weeks lab was about land use and land cover. I was given an aerial photo of Pascagoula, Mississippi, and I was meant to locate features and categorize them based on a land use claddification system. I was able to do this by creating a polygon shapefile and starting and editing session. I then had to manually digitize each polygon and give it a specific code based on what I think that polygon best represented. Codes included were things like residential, commercial, industrial, wetland areas, lakes, forest areas, etc.

Below is the map I map of Pascagoula, Mississippi. Each classification is represented by a different color. I chose colors that would stand out best from the surrounding polygons. Even though this lab took a few hours to complete, I still thought it was fun and interesting. It kind of felt like a puzzle. :)