Module 5: Unsupervised & Supervised Classification

 

Land Use Classification: A Study of Germantown, Maryland

Introduction

Understanding land use is vital for urban planning and environmental management. Using supervised classification techniques in ERDAS Imagine, we analyzed satellite imagery of Germantown, Maryland, to map various land use types and their distribution across the region. The results provide a snapshot of the area's development and natural resources, supporting sustainable growth initiatives like Maryland’s "Smart, Green, and Growing" plan.

Methodology

The classification process began with collecting spectral signatures for eight land use categories:

  1. Urban/Residential
  2. Roads
  3. Grass
  4. Deciduous Forest
  5. Mixed Forest
  6. Fallow Fields
  7. Agriculture
  8. Water

To ensure accuracy, these signatures were refined using tools like the Region Growing Seed Tool. Using a combination of bands (Red: 6, Green: 5, Blue: 4), we classified the entire image with the Maximum Likelihood algorithm, producing a thematic raster map.

A distance image was also generated to identify areas of spectral confusion, guiding adjustments to ensure accurate classification.

Results



The classification revealed the following land use areas:

  • Urban/Residential: 5.096 square miles
  • Roads: 4.625 square miles
  • Grass: 3.231 square miles
  • Deciduous Forest: 11.457 square miles
  • Mixed Forest: 5.216 square miles
  • Fallow Fields: 1.650 square miles
  • Agriculture: 24.990 square miles
  • Water: 1.155 square miles

The total study area is 57.42 square miles, and it is dominated by agricultural land (43.5%) and deciduous forests (20%).

Insights and Implications

  • Urban Development: Residential and road infrastructure accounts for 17% of the area, indicating moderate urbanization.
  • Natural Resources: Forested areas (both deciduous and mixed) cover 29% of the region, highlighting its ecological value.
  • Agricultural Dominance: The significant share of agricultural land emphasizes Germantown's reliance on farming but may point to potential conflicts with urban growth.

This land use map can inform decisions on zoning, resource conservation, and urban expansion to balance development with environmental preservation.

Conclusion

Supervised classification techniques offer powerful insights into land use and land cover. By understanding the spatial distribution of urban, agricultural, and natural areas, we can better address growth challenges while preserving the environment. Germantown’s land use map demonstrates the utility of GIS tools in achieving these goals.

Comments

Post a Comment

Popular posts from this blog

M1 - Python Environments & Flowcharts

Image
  Python Module 1 Recap: From Scripts to Flowcharts and Zen Principles Converting 3 Radians to Degrees To convert 3 radians to degrees, we use a simple formula: degrees = radians * 180 / pi Below is the flowchart illustrating the conversion process: Running the calculations in IDLE gives us:  ≈ 171.887 Interpretation of the Zen of Python The "Zen of Python" showcases the philosophy and guiding principles of Python programming. It emphasizes simplicity, readability, and clarity in code. In other words, “Zen of Python” promotes straightforward and elegant solutions over complicated ones. By valuing explicitness over implicitness, the basis of Python encourages clear and understandable code. Moreover, The principles highlight the importance of practicality and avoiding unnecessary complexity— why use 4 lines of code when 1 will do the same thing? This mentality helps to balance between purity and practicality. Overall, the “Zen of Python” advocates for writing code that is easy ...
Read more

GIS Map Projections

Image
  This week's lab exercise looked at the importance of map projections in GIS, especially when using multiple datasets. The map above shows four counties in Florida that are displayed using the Albers Conical Equal Area, NAD 1983 UTM Zone 16N, and NAD 1983 HARN State Plane Florida North FIPS 0903 (US Feet) map projections. As the table within the map shows, the areas of the counties vary depending on the map projections. The best map projections are ones that shows the least distortion between map features. For example, The UTM projection is best when used the map features fit within one UTM zone, which is not the case for showing the entire state of Florida. The same can be said for the Florida State Plane that is dependent on the region of interest. Therefore, the map projection would need to be Albers since it minimizes distortion between its standard parallels.  Overall, this lab illustrates how important it is to use accurate map projections since it can cause distor...
Read more

ArcGIS Pro Orientation

Image
  Module 1 Lab Assignment This first lab assignment's objective is to prepare and show what is expected for future lab assignments. The primary learning outcomes involved: Getting familiar with accessing the ArcGIS software through Argo Apps. Assessing data storage drives. Preparing project folders for each project. The map shown above is a visualization of the Cities and World Countries data files found in the OverviewArcGIS file in the R drive. In order to view the shapefiles in a way that is informative, we had to add a Folder Connection to the data we downloaded as well as update the symbology to show graduated colors for the field POP_2007 with Natural Breaks and 7 classes. It is important to use a single color from light to dark when using sequential data. While I wish this had been a smooth and uncomplicated process, I struggled with learning UWF's Argo Apps since that is how we access ArcGIS Pro without downloading it to our personal computers. While I am familiar with ...
Read more