Implementation of real time image processing algorithms by using system on chip fpga architecture

In this thesis, gray level image transformation and edge detection algorithms are implemented in real time on Zynq-7000 architecture FPGA. In the experimental setup created, Picozed embedded vision card was used for embedded system application. The card used contains the XC7Z030 FPGA with Xilinx Zynq-7000 architecture. Zynq-7000 architecture is software and hardware programmable because it has ARM processor and FPGA blocks. Communication between these two structures is provided via the Advanced Extensible Interface (AXI). The ARM processor also allows the use of an operating system. The Linux operating system provided by Xilinx provides driver support for the hardware used in the FPGA and the hardware the card contains. In this thesis, the Software Defined System on Chip (SDSoC) programming environment was chosen for the card programming and the Picozed compatible SDSoC image processing platform provided by AVNET was used. In the program, a project that runs on the Linux operating system installed on the card was created using C/C ++ languages. It is also possible to do the same media hardware programming for image processing applications. The hardware calculations are designed directly in the same environment thanks to the hardware accelerated calculation structure. Selected image processing applications; It was realized in 1920X1080 pixel resolution, 60 frames/second in real time.Applied edge detection algorithms are Sobel, Prewitt and Roberts edge detection methods. The image data received from the HDMI input of the embedded system is processed and transferred to the output in the HDMI protocol. In this thesis, each application was realized both in software and hardware and their performances were compared. Keyword: FPGA, Zynq-7000, Picozed embedded vision card, SDSoC, Edge detection