https://github.com/agr-ayush/Landsat-Time-Series-Analysis-for-Multi-Temporal-Land-Cover-Classification

LANDSAT Time Series Analysis for Multi-temporal Land Cover Classification using Machine Learning techniques in Python and GUI development for automation of the process.
https://github.com/agr-ayush/Landsat-Time-Series-Analysis-for-Multi-Temporal-Land-Cover-Classification

land-cover machine-learning python random-forest time-series-analysis tkinter-gui

Last synced: 28 days ago
JSON representation

LANDSAT Time Series Analysis for Multi-temporal Land Cover Classification using Machine Learning techniques in Python and GUI development for automation of the process.

• Host: GitHub
• URL: https://github.com/agr-ayush/Landsat-Time-Series-Analysis-for-Multi-Temporal-Land-Cover-Classification
• Owner: agr-ayush
• Created: 2018-07-11T13:56:14.000Z (over 6 years ago)
• Default Branch: master
• Last Pushed: 2022-02-08T07:56:56.000Z (almost 3 years ago)
• Last Synced: 2024-08-04T04:04:53.398Z (4 months ago)
• Topics: land-cover, machine-learning, python, random-forest, time-series-analysis, tkinter-gui
• Language: Python
• Homepage:
• Size: 13.7 KB
• Stars: 60
• Watchers: 3
• Forks: 32
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

• Awesome-Earth-Artificial-Intelligence - Landsat Time Series Analysis for Multi-Temporal Land Cover Classification

README

        
## Landsat-Time-Series-Analysis-for-Multi-Temporal-Land-Cover-Classification

The Time series analysis for the landsat images was implemented using the random forest machine learning algorithm.

The algorithm classified the image into 4 classes:

1. Vegetation

2. Urban

3. Bare Soil

4. Water

To do the classification different phases were implemented.

1. The first phase includes the calculation of NDVI (Normalized Difference Vegetation Index) and the MNDWI(Modified Normalized Difference Water Index).

2. The second phase involves the stacking of all these .tiff files into a single .tiff file.

   The stacking generally involves the use of bands from 2 to 7, annual NDVI files and the MNDWI file of the selected day.

3. Using this Stacked image we predict the classes using our random forest algorithm and classify the images into the above mentioned 4 classes.