Raw image data previewer

Raw image data previewer is an open-source utility dedicated to parsing and displaying binary data acquired straight from a camera.

This simple utility provides features like:

• previewing most used raw color formats
  • RGB-like formats
  • YUVs (packed, semiplanar, planar)
  • Grayscales
  • Bayer RGBs
• exporting raw image data to more complex formats (ie. PNG, JPG)

Installation

Requirements

• Python >=v3.9
• numpy
• opencv
• PIL

Installation

Manjaro Linux

sudo pacman -Sy python-pip git
git clone https://github.com/antmicro-labs/raw-image-data-previewer.git
cd raw-image-data-previewer
pip install -r requirements.txt

Ubuntu 20.04

sudo apt-get install python3-pip git python3-pil.imagetk
git clone https://github.com/antmicro-labs/raw-image-data-previewer.git
cd raw-image-data-previewer
python3 -m pip install -r requirements.txt

Usage

To start an empty GUI (without any data loaded) use:

cd raw-image-data-previewer
python3 -m app

You can also start the GUI with already loaded data and parameters (like width and color format). More information about available arguments can be found in command-line help:

cd raw-image-data-previewer
python3 -m app --help

Exporting images

The utility also provides a way to convert binary files containing image data to more complex formats (ie. PNG, JPG) without starting the grahpical interface. To use this option simply add the --export argument with the target filename with extension.

Supported formats information

Currently supported color formats and planned ones can be found in the documentation.

Extending supported color formats

Adding new color formats

Currently there are two classes that can describe color formats: ColorFormat and SubsampledColorFormat (found in app/image/color_format.py). To create a new color format, simply:

1. Under AVAILABLE_FORMATS list in color_format.py add new instance of one of the color format classes with proper fields filled.
2. Provide parsing and displaying function by extending AbstractParser found in common.py or by using an existing one.
   • If you choose to implement a new one remember that parse() should return one dimensional ndarray with values read from the binary file. display() on the other hand should return RGB-formatted 3-dimensional ndarray consisting of original color format values converted to RGB24.
3. The utility provides proper parser by checking color format parameters (mainly PixelFormat), so make sure, that your new color format has a valid translation of parameters to one of the parsers (this functionality can be found in app/parser/factory.py).

Contributors

• Błażej Ułanowicz (blazejulanowicz)
• Dawid Dopart (DopartDawid)
• Maciej Tylak (Ty7ak)
• Maciej Franikowski (MaciejFranikowski)

License

The utility is licensed under the Apache-2 license. For details, please read the LICENSE file.