This library primarily provides idiomatic bindings and APIs for OpenCV 3.x.
OpenCV (Open Source Computer Vision Library: http://opencv.org) is an open-source BSD-licensed library that includes several hundreds of computer vision algorithms. It's mainly developed in C++. This library provides Rust bindings to access OpenCV functionalities. First, C bindings are created (in native folder); then Rust APIs are constructed atop. Although this manual process seems an inefficient process, it has served me well as a learning experience to both OpenCV and Rust. In terms of OpenCV API coverage, modules and functions are implemented as needed.
Please check out the documentation to see what has been ported. If you have demand for porting specific features, please open an issue, or better create a PR.
Attempts to use rust-bindgen or cpp_to_rust haven't been very successful (I probably haven't tried hard enough). There is another port opencv-rust which generates OpenCV bindings using a Python script (more automated).
Before anything, make sure you have OpenCV 3 installed. If you are using windows, follow this instruction, otherwise read this Introduction to OpenCV to get started.
Then in any Rust project, add this to your Cargo.toml
:
[dependencies]
cv = { git = "https://github.com/nebgnahz/cv-rs.git" }
And add this to your crate:
extern crate cv;
use cv::*;
And then, enjoy the power of OpenCV.
If you'd like to use OpenCV GPU functions, it's inside cv::cuda
. Enable it
with the following code in Cargo.toml
:
[dependencies.cv]
git = "https://github.com/nebgnahz/cv-rs"
features = [ "cuda" ]
All possible features are listed below:
cuda
- for CUDA support, requires installed CUDAtext
- for text recognition support. Requires building from sources, is not included in most package managers by default, e.g. in brewtesseract
- for Tesseract OCR support, requires installed TesseractC:\opencv
..git
and .windows
folders there (you can run them from the cv-rs
directory itself, but you may encounter an error that paths are too long)c:\opencv
.local
package.PowerShell -NoExit -File .\.windows\msvc_1_install_CUDA.ps1 -FileName path_to_installer
(for example, C:\Users\UserName\Downloads\cuda_9.1.85_win10.exe
).PowerShell -NoExit -File (.\.windows\msvc_2_build_OCV.ps1 -EnableCuda $False -Compiler vc15)
(note braces). 1
stays for compilation with CUDA, 0
for compilation without it. Possible compiler values: vc14
for VS2015/vc15
for VS2017. Caution: CUDA is compatible with VS2015 onlyx86_64
during installation.C:\opencv
..git
and .windows
folders there (you can run them from the cv-rs
directory itself, but you may encounter an error that paths are too long)c:\opencv
.PowerShell -NoExit -File .\.windows\mingw_build_OCV.ps1 -MinGWPath "C:\Program Files\mingw-w64\x86_64-7.2.0-posix-seh-rt_v5-rev1\mingw64\bin"
(your path may be different).See available examples on how this library might be used.
See the contribute file! PRs highly welcome.
You may also simply open up an issue for feature/porting request.
Small note: If editing the README, please conform to the standard-readme specification.
MIT © Ben Zhang