Tuesday, November 14, 2017

GIS 4035: Module 10 - Supervised Classification


Above is a map of Germantown, MD's Land Use. In this lab we learned how to further classify images in ERDAS by creating AOI layers that allow us to create our own signatures. By using the polygon and grow properties tool, we were able to pinpoint certain coordinates and areas and classify them in the image. After re-coding the signatures, we were left with 8 unique classifications that are displayed on the map. 

Tuesday, November 7, 2017

Module 9 - Unsupervised Classification


This week we learned how to perform an unsupervised classification in both ArcMap and ERDAS. We we were able to use tools and techniques such as iso cluster, maximum likelihood classification tool, swiping,  blending, flickering, and recoding. To demonstrate our new skills, we used an image of the UWF campus to classify 5 different area types.  The image above shows the breakdown of the 5 new categories and the percentage of acreage they cover in the land.

Tuesday, October 31, 2017

Module 8: Thermal & Multispectral Analysis


This week we continued to build on our photo analysis skills. The focus was thermal and multispectral analysis. We were able to create composite images in both ERDAS and Arcmap, interpret thermal infrared data, and we were able to continue working on band combinations. For this particular lab we had to choose one feature to display within the image provided. The bridge that decided to feature was displayed with a band of Red (Layer 6), Green (Layer 5), and Blue (Layer 4). You could notice the bridge going across the rivers. The small strip of pavement must provide enough thermal heat next to the water to create a great visual.

Tuesday, October 24, 2017

GIS 4035 Module 7: Multispectral Analysis

This week we learned several new techniques when it comes to multispectral analysis. We learned histograms, spectral characteristics, using the unique cursor, and using gray scale. We then tied all of our skills together to determine the best band combinations.
 The purpose of this lab was to determine 3 features using different band combinations.  Each map listed below has the explanation of each image and band combination. This lab was very interesting and I enjoyed learning how to change the image quality to display certain features.




Tuesday, October 17, 2017

Module 6: Spatial Enhancement


This weeks lab allowed us to get creative and start using the skills that we had learned. We were able to use newly acquired tools to make corrections to images that may have imperfections or errors. We were able to use both Erdas and Arcmap to utilize tools to enhance an image. One of the new tools that we learned how to use was the Fourier Transformation. We were able to use that to take the major striping errors out of the image. We also used high pass and low pass filters to make the image clearer. I was able to experiment a lot in this lab and learned many new skills that I will be able to utilize in the future.

Tuesday, October 3, 2017

Lab 5a: Intro to ERDAS Imagine and Digital Data 1


This week we began with some calculations of in relations to wavelength, frequency, and energy of EMR. I was also introduced to Erdass Imagine. The lab walked us through some of the basics of the program. We used the new program to create a subset image or map using a classified image of a section of forested lands in Washington State. I found some similarities of the program that made it easy to adjust. We then uploaded the subset image to ArcMap to complete our map. I found this week very interesting and I am looking forward to building on what I learned.

Tuesday, September 26, 2017

Lab 4: LU/LC Ground Truthing Analysis

This week we were able to continue off of last modules lab. Last week we worked on LU/LC classification. This week we had to check the accuracy of the lab using ground truthing analysis. After placing 30 points throughout the map, we used google maps to determine whether our initial analysis was true or false. Out of the 30 points, 27 of them were found to be true resulting in 90% accuracy.