Bayesian_River_Classification

This project implements a Bayesian decision rule for classifying River and Non-river areas using satellite images.

Dataset

The dataset consists of four satellite images in .gif form which can be downloaded from here.

Requirements

• Python 3.x
• numpy
• pandas
• matplotlib
• scikit-image
• scikit-learn
• seaborn
• opencv-python

Installation

pip install numpy pandas matplotlib scikit-image scikit-learn seaborn opencv-python

Usage

1. Annotate Points:

Run the following script to manually annotate 50 points from the river and 150 points from non-river in band4 image:

python annotate_image_points.py

2. Run Classification:

Run the band_image_read.ipynb to classify the River and Non-river areas and plot the confusion matrix:

Scripts

• annotate_image_points.py Script for annotating points on the image.
• band_image_read.ipynb Script for classifying River and Non-river areas and plotting confusion matrix.
• Results The output will be a classified image saved as image.png and a confusion matrix plot.