igorm::home‎ > ‎igorm::projects‎ > ‎

igorm::projects / objects recognition

 home | projects | drawings | ideas

Objects Recognition

Personal, Research

The program reads the video stream from a WebCam (connected to PC). It does all the required image processing and tries to recognise the objects placed in the view of camera. At the moment the system can identify the simple objects (rectangles, triangles, circles, polygons). The outcome of the system is a list of objects and their properties (including vectorial representation). The program consists in a set of execution threads, placed in a cascade form: first reads the video stream, second processes the frames, the last one identifies the objects.
The camera is providing a 10FPS video stream. The processing is done in real time.

Linux / GC++ / V4L / WebCam


Current status
Under development / A functional prototype is available.

Detailed description
The following video displays a demo of the running program: