This project implements a cross-platform tool for focus stacking images. The application takes a set of images captured at different focus distances and combines them so that the complete subject is in focus.
Binary packages for Windows 10, Ubuntu 20.04 and Mac OS X are available on the GitHub releases tab.
Note that Mac OS X may end up putting the application to quarantine because it is unsigned. This gives misleading error message "focus-stack.app” is damaged and can’t be opened. You should move it to the Trash." You can try this command to remove the protection:
xattr -d com.apple.quarantine focus-stack.app
In most cases just passing all the images is enough:
build/focus-stack .../path/to/input/*.jpg
Input images can be in any format supported by OpenCV, including JPG, PNG and TIFF.
Output format defaults to JPG, but can be set by changing the extension with --output=output.png.
For advanced usage, see --help for list of all options or check the manual:
Usage: build/focus-stack [options] file1.jpg file2.jpg ...
Output file options:
--output=output.jpg Set output filename
--depthmap=depthmap.png Write a depth map image (default disabled)
--3dview=3dview.png Write a 3D preview image (default disabled)
--save-steps Save intermediate images from processing steps
--jpgquality=95 Quality for saving in JPG format (0-100, default 95)
--nocrop Save full image, including extrapolated border data
Image alignment options:
--reference=0 Set index of image used as alignment reference (default middle one)
--global-align Align directly against reference (default with neighbour image)
--full-resolution-align Use full resolution images in alignment (default max 2048 px)
--no-whitebalance Don't attempt to correct white balance differences
--no-contrast Don't attempt to correct contrast and exposure differences
--align-only Only align the input image stack and exit
--align-keep-size Keep original image size by not cropping alignment borders
Image merge options:
--consistency=2 Neighbour pixel consistency filter level 0..2 (default 2)
--denoise=1.0 Merged image denoise level (default 1.0)
Depth map generation options:
--depthmap-threshold=10 Threshold to accept depth points (0-255, default 10)
--depthmap-smooth-xy=20 Smoothing of depthmap in X and Y directions (default 20)
--depthmap-smooth-z=40 Smoothing of depthmap in Z direction (default 40)
--remove-bg=0 Positive value removes black background, negative white
--halo-radius=20 Radius of halo effects to remove from depthmap
--3dviewpoint=x:y:z:zscale Viewpoint for 3D view (default 1:1:1:2)
Performance options:
--threads=2 Select number of threads to use (default number of CPUs + 1)
--batchsize=8 Images per merge batch (default 8)
--no-opencl Disable OpenCL GPU acceleration (default enabled)
--wait-images=0.0 Wait for image files to appear (allows simultaneous capture and processing)
Information options:
--verbose Verbose output from steps
--version Show application version number
--opencv-version Show OpenCV library version and build info
On Windows you can additionally just select the photos and drag them
over focus-stack.exe to run with default settings.
On Mac OS X, you can start the program directly and it will ask for files.
Alternatively you can drag the files over focus-stack application.
If you want to provide command line parameters, open terminal and call the
binary with path focus-stack.app/Contents/MacOS/focus-stack.
With default settings, approximately 100 MB of RAM is required per 1 megapixel of input image resolution. The images are processed in batches, so total number of images in stack does not significantly affect memory usage.
Memory usage can be reduced at the cost of computation speed.
It is affected by number of images per batch, and number of simultaneous
processing threads.
Minimal configuration of --threads=1 --batchsize=2 uses about 50 MB per megapixel.
The focus stacking algorithm used was invented and first described in Complex Wavelets for Extended Depth-of-Field: A New Method for the Fusion of Multichannel Microscopy Images by B. Forster, D. Van De Ville, J. Berent, D. Sage and M. Unser.
The application also uses multiple algorithms from OpenCV library. Most importantly, findTransformECC is used to align the source images.
For more information, see the detailed explanation of algorithms.
The only dependency is OpenCV, 3.0 or newer, plus the basic build tools:
sudo apt install libopencv-dev build-essential # Ubuntu
sudo dnf install opencv-devel make gcc-c++ # Fedora
To enable GPU acceleration, you additionally need OpenCL library and GPU-specific driver:
sudo apt install ocl-icd-opencl-dev beignet # For Intel GPUs
sudo apt install ocl-icd-opencl-dev nvidia-opencl-icd # For NVidia GPUs
To build and install the application, simply type:
make
make install
Or to build a Debian/Ubuntu package and install it, type:
sudo apt install debhelper devscripts
make builddeb
sudo dpkg -i DEBUILD/focus-stack*.deb
Download OpenCV binary package for Windows from OpenCV website. Download Build Tools for Visual Studio 2019 from Microsoft.
In Visual Studio command prompt, execute:
nmake -f Makefile.windows
Install Homebrew by running this in terminal:
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install.sh)"
Then install build dependencies:
brew install opencv pkg-config dylibbundler
Finally, build and install to /usr/local/bin:
make
make install
Alternatively you can build an application bundle that can be used on any machine and that also includes a simple GUI for selecting the files:
make build/focus-stack.app
There is a third-party Android app that uses focus-stack for algorithm:
