CVOCV--iOS- is a project mainly written in OBJECTIVE-C and C++, it's free.
An OpenCV-based video experimentation environment for iOS 4.x.
CVOCV was built in order to create an extremely simple and lightweight OpenCV experimentation environment for OS X 10.5.x and iOS 4.x. Using CoreVideo makes it possible to send video frames to OpenCV with a minimal amount of computational overhead.
The current implementation was designed to target the second generation iPad (iPad2). Targeting the iPhone running iOS 4.x shouldn't pose any significant issue, should you decide to target that device instead.
In order to get this project to run on an iOS device, you'll first need to build OpenCV. These instructions worked for me, and I highly recommend following them to build yourself a version of OpenCV that you can run on your iOS device.
http://computer-vision-talks.com/2011/02/building-opencv-for-iphone-in-one-click/
Once you're done building, you'll need to change some of the project settings to reflect where you've built OpenCV. Follow these instructions to do so (Note: you'll only need to follow the first two steps to set up your include and lib directories:
http://computer-vision-talks.com/2011/01/using-opencv-in-objective-c-code/
At the heart of CVOCV is the 'captureOutput' method found in CVOCV_IOSViewController.mm. This is a callback procedure that is invoked by CoreVideo to provide direct access to the pixels coming from a video input device.
To experiment with the existing examples offered by CVOCV, simply uncomment one of the lines in the VideoProcessor::setFrameImage() method.
For example, specifying the line
return adaptiveThreshold(frameImage);
at the end of setFrameImage() will perform adaptive thresholding on the incoming video frame and display the result on screen.
Feel free to add your own examples and try them out!
There are three native capture sizes for video frames on iOS devices: (192x144, 640x480, 1280x720). You may toggle between these sizes via the 'captureSize' variable defined in Constants.c. Note that increasing the capture frame size will increase the computational overhead of the OpenCV processing step.
You may switch between which camera device is used by editing the 'startCamera' method in CVOCV_IOSViewController.mm.
Contributions in the form of elegant, demonstrative examples are welcome and encouraged. Contact me if you have something you think would make a nice addition to this project.
2011 Kyle Buza [email protected]