Development of a Robot for Identifying a Unique Object

Abstract

Identifying unique object and separating it from other objects is a major task in most industries. The objects can be separated based on colour and shape parameters. In this paper a robot is designed and tested to identify unique object based on shape in a given specified region. It assumed that the object may be a cube or a sphere. A unique object identified will have either different colour or different shape when compared to the remaining object. The robot keeps moving in a specified region, when an obstacle is detected it stops and captures an image through camera, which is mounted on robot. This captured image is given to Beagleboard where it performs image processing task to identify type of object. The Beagleboard used to perform image processing task comes with ARM@ CortexTM-A8 processor with 512 MB memory and low-power DDR RAM. The robot takes less than 10 ms of time to identify an object.