changedet

A Python toolkit for classical change detection algorithms in remote sensing and geospatial imagery.

Features

changedet implements 4 classical change detection algorithms:

• Image Differencing (imgdiff) - Simple pixel-wise difference
• Change Vector Analysis (cva) - Magnitude and direction of change with Otsu thresholding
• Iterated PCA (ipca) - Principal Component Analysis with iterative reweighting
• IR-MAD (irmad) - Iteratively Reweighted Multivariate Alteration Detection

All algorithms work with multi-temporal satellite/aerial imagery in GeoTIFF format.

Installation

From source

git clone https://github.com/ashnair1/changedet
cd changedet
pip install .

For development

git clone https://github.com/ashnair1/changedet
cd changedet
poetry install --with dev,docs

Quick Start

List available algorithms

changedet list

Run change detection

# Basic usage
changedet --algo imgdiff run image1.tif image2.tif

# Change Vector Analysis with Euclidean distance
changedet --algo cva run image1.tif image2.tif --distance euclidean

# IR-MAD with 10 iterations
changedet --algo irmad run image1.tif image2.tif --niter 10

Get help

# General help
changedet --help

# Algorithm-specific help
changedet --algo cva algo_obj --help

Documentation

Full documentation is available at https://ashnair1.github.io/changedet/

License

MIT License - see LICENSE file for details.