OpenCV.jl

The Julia wrapper for Open Source Computer Vision library (OpenCV)

https://github.com/JuliaOpenCV/OpenCV.jl

OpenCV — Module.

The Julia wrapper for Open Source Computer Vision library (OpenCV).

A quick demo:

julia> using CVCore, CVImgCodecs, CVHighGUI

# Reading a image file into a Mat (equivalent to cv::Mat in C++)
julia> A = imread(Pkg.dir("OpenCV", "test", "lena.png"))
CVCore.Mat{UInt8,3}
[:, :, 1] =
 0x4a  0x4f  0x4c  0x4b  0x50  0x50  …  0x37  0x37  0x36  0x37  0x37  0x37
 0x4d  0x4e  0x4b  0x4d  0x52  0x4c     0x3a  0x3a  0x39  0x39  0x39  0x39
 0x4e  0x4c  0x4d  0x4f  0x4b  0x4b     0x36  0x32  0x34  0x36  0x36  0x37
 0x50  0x48  0x52  0x4e  0x4a  0x50     0x35  0x35  0x38  0x39  0x3a  0x37
 0x52  0x46  0x56  0x4d  0x49  0x53     0x35  0x36  0x39  0x3b  0x3b  0x3a
 0x52  0x49  0x53  0x4e  0x4a  0x50  …  0x37  0x34  0x37  0x3b  0x3d  0x3d
 0x4e  0x51  0x4a  0x51  0x4d  0x47     0x39  0x36  0x38  0x3d  0x3d  0x3d
 0x4c  0x52  0x46  0x50  0x4d  0x44     0x3a  0x3a  0x3e  0x3d  0x3a  0x38
 0x4b  0x4d  0x48  0x4d  0x4a  0x47     0x39  0x3b  0x3b  0x3a  0x33  0x2e
    ⋮                             ⋮  ⋱                             ⋮

# You can convert it to a Julia array
julia> B = Array(A)
3×512×512 Array{UInt8,3}:
[:, :, 1] =
 0x4a  0x4f  0x4c  0x4b  0x50  0x50  …  0x37  0x37  0x36  0x37  0x37  0x37
 0x6f  0x74  0x71  0x6f  0x74  0x73     0x4e  0x4e  0x4d  0x4e  0x4e  0x4e
 0xc1  0xc6  0xc3  0xc3  0xc8  0xc9     0x92  0x92  0x91  0x92  0x92  0x92

[:, :, 2] =
 0x4d  0x4e  0x4b  0x4d  0x52  0x4c  …  0x3a  0x3a  0x39  0x39  0x39  0x39
 0x72  0x73  0x70  0x71  0x76  0x6f     0x51  0x51  0x50  0x50  0x50  0x50
 0xc4  0xc5  0xc2  0xc5  0xca  0xc5     0x95  0x95  0x94  0x94  0x94  0x94

[:, :, 3] =
 0x4e  0x4c  0x4d  0x4f  0x4b  0x4b  …  0x36  0x32  0x34  0x36  0x36  0x37
 0x73  0x6b  0x75  0x71  0x6d  0x73     0x4e  0x4c  0x4e  0x50  0x50  0x4f
 0xc7  0xc0  0xc8  0xc5  0xc1  0xc6     0x90  0x92  0x94  0x96  0x93  0x91

# Show the image
julia> imshow("lena", A); waitKey(0) # you will see a GUI window

Dependencies (in short)

General API guidelines

Function names should be same between Julia and C++, except for appending ! (e.g. threshold!), which means the fuction does in-place operations.
cv::Mat in C++ are represented as Mat{T,N} in Julia (see CVCore.jl for details).

Package structure

To simplify development and minimize dependencies, the Julia wrapper consists of the packages below:

following the opencv module structure except for the LibOpenCV.jl which manages binary dependencies (i.e. search installed OpenCV shared libraries or build and install them if not found).

source