s-du / skyspectral

A simple tool to process multispectral images from the Micasense RedEdge-P (including indices, Agisoft, etc.)

SkySpectral

A simple tool to process multispectral images from the Micasense RedEdge-P

Coming next:

• Pan sharpening algorithm
• Agisoft automated workflow
• Image annotation

Introduction

SkySpectral is a simple tool for visualizing and processing multispectral imagery. It enables the alignment and combination of various sub-images (bands) to extract relevant data.

#Multispectral #Open-CV #Building diagnosis #ImageProcessing

The project is still in pre-release, so do not hesitate to send your recommendations or the bugs you encountered!

Processing multispectral image, from a folder

Principle

The functionalities include:

• Import Micasense output folder
• Visualize bands with standard matplotlib color palettes
• Align channels (R,G,B,NIR,Red-Edge) to Pan-chromatic with an intuitive interface (especially critical for close-shots, as in building façades analyses)
• Visualize composed shot (RGB image re-composed from individual channels, CIR - Color Infrared, etc.)
• Create custom indices based on arithmetic operations on bands
• Access typical vegetation indices such as NDVI
• Prepare the files for Agisoft Metashape (organize images in folders)
• ...

Creating custom indices

Installation instructions

1. Clone the repository:

git clone https://github.com/s-du/SkySpectral

2. Navigate to the app directory:

cd SkySpectral

3. Install the required dependencies:

pip install -r requirements.txt

4. Run the app:

python main.py

Contributing

Contributions to the app are welcome! If you find any bugs, have suggestions for new features, or would like to contribute enhancements, please follow these steps:

1. Fork the repository.
2. Create a new branch for your feature or bug fix.
3. Make the necessary changes and commit them.
4. Push your changes to your fork.
5. Submit a pull request describing your changes.