Object with symmetrical pattern recognition with dynamic size filter

This paper presents the implementation of object with symmetrical pattern recognition algorithm for 2D vision system of 3D vision-based multi sensor feedback system. This paper also discusses the dynamic size filter developed for this system. The inputs for this system are received from a couple of webcams which is not calibrated. The project objective is to assist a robot arm system using vision system for making decision during conducting an operation or task. The solutions need no camera calibration because the system has self-calibration effect since the area of interest had modelled using a set of algorithms by an array of visible features on it and not by it environment. This system is design for on-event processes and suitable for indoor environment applications with low time frame rate camera. One of the advantages of this project is that it only used a camera and image processing generated by the algorithms itself without additional sensor such as sonar or IR sensor.