3D reconstruction with Structure Light base on OpenCV
Congratulations for finding this repository, people as smart as you might notice that this repository has cleanner, easier to use structure than others.
We provide noncontact 3D surface measurement techniques using OpenCV.
test images in "res/imgs" folder are compressed
I cannot promise the quality of the point-cloud from compressed images, so the testdata is just used for
academic purpose. if you pursue high quality of the point-cloud, you should take the images with high resolution.
If you are interested in creating picture from camera with projector, you can have a look at my hikvision-qt.
The Projected Patterns contains gray codes and phase shift
The Calibration process uses Chessboard for canmera parameters.
(It supposed to have Dot array calibration pattern... but I didn't have the board.)
To use the code you need to install OpenCV3(only tested in version 3.xx), and also you need to install CMake to configure the code.
git clone https://github.com/timbrist/structure-light.gitcd structure-lightmkdir buildcd buildcmake ..make -j4./main
Put your images into the folder "../res/imgs/origin/",
Name the images as "./res/origin_images.xml" suggested.
This is the 3d point cloud:
Open the the "point.txt" with your favorite point cloud software.
here is the result that shows in " geomagic studio 2015 "
