Abstract:
Image Identification has become very important due to its applications in the
areas of automated quality assurance systems, industrial processing, product sorting and
fault identification. It involves 2D signal processing that is a multi-step process
comprising of image acquisition, data processing, image classification & analysis and
information output.
The system designed in this project acquires a two dimensional image. First of all
the image is segmented and analyzed for its features like edges and distance transforms
The features of the image are then assigned appropriate weights and the resultant is
matched with an existing template database. The matching is invariant to rotational and
translational shifts, and to scaling. If the best match is above a designated criterion, the
object is recognized (and the object is said to exist in the database), otherwise it is treated
to be a new object. In this way the software has the provision to automatically update its
existing database for a new object. To make it a real-time system, the optimized code is
written in Visual C++. Furthermore, the function accelerator is designed for gradient
operation to show the hardware design of the algorithm. Gradient operation is a
computationally intensive process so the hardware has been designed for it. A graphical
user interface (GUI) is designed to monitor the intermediate results of processing which
makes the system even more user-friendly.
