#bayer #cfa #demosaic


Algorithms for demosaicing Bayer images

6 releases

Uses old Rust 2015

0.1.5 Jul 22, 2017
0.1.4 Apr 29, 2017
0.1.3 Feb 17, 2017

#154 in Images

Download history 951/week @ 2022-04-21 596/week @ 2022-04-28 612/week @ 2022-05-05 823/week @ 2022-05-12 949/week @ 2022-05-19 919/week @ 2022-05-26 585/week @ 2022-06-02 762/week @ 2022-06-09 1405/week @ 2022-06-16 608/week @ 2022-06-23 641/week @ 2022-06-30 643/week @ 2022-07-07 523/week @ 2022-07-14 669/week @ 2022-07-21 1121/week @ 2022-07-28 1350/week @ 2022-08-04

3,761 downloads per month
Used in 2 crates (via aravis)

MIT license

1.5K SLoC

LibBayer Version Status


LibBayer provides routines for demosaicing Bayer (raw) images.

The library supports 8-bit and 16-bit images.

Several demosaicing algorithms are available. See the src/demosaic directory for a list and their individual descriptions. Pixels on the border of the image are retained by replicating or mirroring the data in the neighbourhood.

LibBayer is written entirely in Rust. C bindings to the underlying algorithms are provided.


An example program is provided in the examples/ directory:

  • showbayer - a simple Bayer file viewer.
  • writebayer - converts an image to a raw Bayer image file.

To clone this repository, run:

git clone https://github.com/wangds/libbayer.git

Then build the library and run the example programs using Cargo.

cargo build --release --example showbayer

To display a Bayer file, run:

cargo run --release --example showbayer <width> <height> <depth> <example.raw>

Change the colour filter array (CFA) pattern and the demosaicing algorithm from inside the example program.

Basic Usage

Add LibBayer as a dependency to your project's Cargo.toml:

bayer = "0.1"

Import the library in your project, e.g.:

extern crate bayer;

Open a Bayer file from disk.

let mut file = File::open(Path::new("example.raw")).unwrap();

This raw data contains the red, green, or blue values at each pixel only. There is no additional header data, so the image width and height, pixel depth, and CFA pattern must be provided from elsewhere.

Allocate the buffer to which we will decode the image.

let img_w = 320;
let img_h = 200;
let depth = bayer::RasterDepth::Depth8;
let bytes_per_pixel = 3;
let mut buf = vec![0; bytes_per_pixel * img_w * img_h];
let mut dst = bayer::RasterMut::new(img_w, img_h, depth, &mut buf);

Then run the demosaicing process:

let cfa = bayer::CFA::RGGB;
let alg = bayer::Demosaic::Linear;
bayer::run_demosaic(&mut file, bayer::BayerDepth:Depth8, cfa, alg, &mut dst);

Note that many cameras will capture 12-bits per pixel (channel), but store the data as 16-bits per pixel. These should be treated as 16-bits per pixel for the purposes of this library.



David Wang


~30K SLoC