CityChronicle-Bengaluru-LandUse-LandChangeAnalysis

Addressed class imbalance in urbanization study of Bangalore. Applied SVM (86.5%) and Random Forest (95.4%) for change detection (2014-2020). Leveraged pixel differencing, SSIM, and MSE for accurate land cover classification.

In this project, the challenge of class imbalance is adressed while conducting a comprehensive analysis of Land Use and Land Cover Change detection in Bangalore. Focused on monitoring rapid urbanization, the study spans the years 2014, 2016, and 2020. Leveraging state-of-the-art machine learning models such as SVM (86.5% accuracy) and Random Forest Classifier (95.4% accuracy), I successfully tracked and predicted urbanization patterns.

To ensure the precision of our findings, implemented a robust evaluation framework employing image pixel differencing, SSIM, and MSE metrics. These metrics not only provided accurate classification results but also facilitated effective detection of changes in land cover. This project contributes valuable insights into the dynamic urban landscape of Bangalore, combining sophisticated models with meticulous evaluation methods.

LULC & Change Detection for Bangalore (2014 & 2016 & 2020)

Overview

This project leverages Google Earth Engine and Landsat-8 satellite data for Land Use Land Cover (LULC) analysis and Change Detection in Bangalore.

Folders Structure

• 2014_lulc: Contains code for 2014 LULC analysis.
• 2016_lulc: Contains code for 2016 LULC analysis.
• 2020_lulc: Contains code for 2020 LULC analysis.

Instructions

1. Choose the appropriate year's folder (e.g., 2014_lulc).
2. Open code.txt inside the folder.
3. Copy the JavaScript code from code.txt.

Running the Code

1. Open Google Earth Engine's script area.
2. Paste the copied code into the script area.
3. Run the script to generate the output.

Step-by-Step Guide

1. Locate the Code File:

   • Navigate to the project folder.
   • Choose the folder corresponding to the desired year (e.g., 2014_lulc).

2. Open code.txt:

   • Inside the selected folder, find the file named code.txt.
   • Open this file using a text editor.

3. Copy the JavaScript Code:

   • Select all the code within code.txt.
   • Copy the code to your clipboard.

4. Open Google Earth Engine:

   • Visit the Google Earth Engine platform in your web browser.
   • Ensure you are logged in with your Google account.

5. Access the Script Area:

   • On the left panel, click on the "Script" tab to access the script area.

6. Paste the Copied Code:

   • Inside the script area, paste the copied JavaScript code.
   • Review the code for any specific parameters that may need customization.

7. Run the Script:

   • Click on the "Run" button within the script area.
   • This will execute the code in the Google Earth Engine environment.

8. Monitor the Console:

   • Keep an eye on the console for any errors or progress messages.
   • The console will display information about the script execution.

9. Wait for Output:

   • Depending on the complexity of the analysis, the script may take some time to run.
   • Wait patiently for the process to complete.

10. View the Output:

• Once the script finishes running, visualize the output directly in Google Earth Engine.
• Explore the layers and maps generated by the script to observe the LULC analysis or change detection results.

Remember to follow Google Earth Engine's documentation and guidelines for smooth script execution and troubleshooting if needed.

For more details, refer to the attached report.