I want to see the change in various features like vegetation, water bodies, urban area, wetland etc. I have Landsat images. How do I perform this change detection using Python?
There are several ways to do so. The simplest one would be to classify the two images and then compare them so that the output is binary in nature (same land use or not). Gdal and numpy are useful libraries to achieve this.
For the purpose of classification, use machine learning algorithm where you provide a trained spectral signature or index value for each land use type and then perform supervised classification (using, for example, random forest algorithm).
There are several ways to do so. The simplest one would be to classify the two images and then compare them so that the output is binary in nature (same land use or not). Gdal and numpy are useful libraries to achieve this.
For the purpose of classification, use machine learning algorithm where you provide a trained spectral signature or index value for each land use type and then perform supervised classification (using, for example, random forest algorithm).
I would just like to point out two very interesting, related Kaggle competitions:
https://www.kaggle.com/c/dstl-satellite-imagery-feature-detection
https://www.kaggle.com/c/draper-satellite-image-chronology
What are the requirements for Land Use Land Cover Classification in terms of satellite data? Is it required to have the satellite image atmospherically corrected one?
Hey, I saw a presentation today on your specific problem. It was in the context of a challenge organized at ECML. It works with Landsat images ;)
https://sites.google.com/site/dinoienco/tiselc
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Programming Languages
- Python