Processing Computer Vision Library – diewald_CV_kit
main project page (download):
this library contains tools that are used in the field of computer vision.
its not a wrapper of openCV or some other libraries , so maybe you are missing some features ( … which may be implemented in the future).
its designed to be very fast to use it for realtime applications (webcam-tracking, kinect-tracking, …).
also, it works very well in combination with the kinect-library ( dLibs.freenect ) … which i basically built it for … to track blobs, generate contours from 3d-data, and else.
it should work in combination with other libraries too, since the blobdetection is very flexible and works with any given data-arrays.
the examples, that come with the library, demonstrates:
- kinect 3D/2D tracking (requires dLibs.freenect v01.10 or higher)
- a simple marker tracking
- image-blob tracking
- processing examples
- library: diewald_CV_kit.jar
- connected component labeling – blobdetection
a blob has only one outer contours, and can have endless inner contours
the outer contour always goes in clockwise-direction, the inner ones go counter-clockwise
- convex hull
- double-linked-list (used for the covex hull for quick node adding/removing)
- polyline tools (simplification, areasize, length, etc.
- color class for fast generating/extracting of color components
kinect tracking example made with dlibs.freenect.the blobs are detected by analysing the 3d pointcloud.
another kinect-example using “dlibs.freenect”.
the markers are simply detected by comparing the number of contours. but its easily piossible to check for more features on other use-defined markers, to get orientation, size, etc. too.