Stereo vision applying OpenCV and raspberry Pi

Abstract

This article points out the single board computer Raspberry Pi and the related camera modules for image acquisition. Particular attention is directed to stereoscopic image recording and post processing software applying OpenCV. A design of a camera network is created and applied to a field application. The OpenCV computer vision library and its Python binding provides some script samples to encourage users developing their own custom tailored scripts. Stereoscopic recording is intended for extended base lines without a mechanical bar. Image series will be taken in order to wipe out moving objects from the frames. And finally the NoIR camera made infrared photography possible with low effort. Computer, accupack and lens board are assembled in a 3D printed housing operated by a mobile device.