Brooke's GIS Blog

  The main struggle I had was the classification of the roads, urban and buildings. The spectral values I selected were too similar, so the colors blended together when I merged classes. Because I was short on time, I opted to categorize roads, urban and residential as one. I also had trouble adding signatures to the deliverable map, as well as creating a subset. My subset originally came out extremely pixely with no clear boundaries. To fix this, I changed the file to 16 bit signature. Ultimately, if I had managed my time better I would have been able to categorize the roads, urban and residential separately. My excuse was I was working on my Open Water Certificate the last two weekends, and am working on this as I’m doing Thanksgiving travel 😊  

 The SLOs of this lab were to download and import satellite imagery, perform spatial enhancements in ArcgGIS Pro and ERDAS Imagine, explore Image Histograms, Operate the inquire cursor, interpret histogram data in images and create spectral band indices. I enjoyed this lab, I feel like changing spectral band combinations and performing 3x3 low pass/high pass/sharpen is similar to working in Adobe Photoshop. Before this lab, I never knew how large of a difference simple spectral band combination changes makes to an image. The one thing I really struggled with was reading histograms. I just used the inquire cursor, I will need more practice with histograms to get it down.

 The objectives for Lab 3a were to become familiar with ERDAS Imagine and how to subset data as a preprocessing step for creating a map in ArcGIS. I enjoyed this lab, especially learning the basics of  new software. Working with an image that has significant elevational differences was helpful to see the ways in which color bands differ. For example, when switching between TM False Natural Color and False Color IR, I was able to better see differences with terrain layout and even ice capped mountains. Overall, I'm excited to know the basics of ERDAS Imagine and look forward to learning more about it.