Design and Implementation of Image Processing on Epga: Implementation and Optimization of Image Processing Algorithms on EPGA-based Devices

Abstract

The criticality in handling the graphics data in an EPGAdevice has been addressed in this work. We have successfully shown how imagedata can be successfully converted to raw binary data and can be transferred asa file to the EPGA memory using the RS-232 link. Our design is very useful forthose EPGA devices (low cost) which don’t have a dedicated input graphics portbut still there is a need to perform the processing of image data. Moreover, wehave detailed out the EDK development platform for building the embeddedapplications in EPGA systems using Micro blaze soft-core processor. In thiswork we have also given a brief overview of creating programs using System Clanguage. This work can work as a guide for the researchers and engineers whowish to build an embedded system for image processing applications. This paperpresents the design and implementation of image processing applications onfield programmable gate array (EPGA). To improve the implementation time,Xilinx AccelDSP, software for generating hardware description language (HDL)from a high-level MATLAB description has been used. Two EPGA-basedarchitectures for image processing have been proposed: Color Space Conversionand Edge Detection. The designs were implemented on Spartan 3A DSP and Vertex 5devices. Obtained results are discussed and compared with others architectures.Nowadays, image processing applications are frequently preferred in such fieldsas industrial automation, security, health, and traffic control in parallelwith the developments in technology. The most important criterion for theapplications used in these fields is to ensure the system to run at high speedand real time. Thus, EPGAs are commonly used in such applications. In thisstudy, some of the basic real time images processing algorithms are implementedin an EPGA-based development kit.