Visual Tracking Software

Recently there has been much interest in computer vision and tracking from a live or recorded video stream. Some application include tracking of people using facial recognition or tracking object motion. More recently, a lot of the tools and libraries are now open source.

The most popular is OpenCV (Open Source Computer Vision), developed by Intel Russia research center. With this tool, you can grab any camera, wecam, or laptop camera, and be able to track things easily. For example, if you open up OpenCV and execute “facedetect.exe” you can see it track your face.

Another user friendly tool is OpenTLD, originally developed in Matlab by Zdenek Kalal for his thesis. What makes this software great is that it does not use any training or a priori data for the tracking. There is also an C++ implementation. Open TLD is no longer supported and more recently, CMT (Consensus-based Matching and Tracking) has emerged and is being developed.

From an aerospace or mechanical engineering view, this would be a very useful tool to track an object and obtain it’s position and velocity (by finite differencing, etc.) as it moves. For example, looking an object falling or a motion of a particle as it moves through a medium.

Instruction of OpenTLD (Matlab)
OpenTLD source files (Matlab)

Example of OpenLTD

Example of CMT

Posted in Uncategorized

Leave a Reply

Your email address will not be published. Required fields are marked *