Porovnání efektivity zpracování obrazu na různých HW platformách

Abstract

The main task of this bachelor thesis is the comparison of three hardware platforms based on the efficiency of image processing. Efficiency is determined over the length of the processing time. Two platforms myRIO and Industrial Controller use a FPGA chip for fast data processing together with a processor. The third platform is a computer. This is a comparison of a classical CPU with a FPGA chip, and in the case of an FPGA chip, the data have to go from the side where the processor runs on the FPGA side, which can be a narrow throat of the whole processing. The chosen image processing algorithm is the Canny edge detector, which is computationally enough to identify the performance deficiencies of individual platforms. The chosen algorithm is theoretically described, which is followed by the design and implementation of the tests. The result of the thesis is a table with image processing times, which compares also with the LabVIEW library function. There is also a theoretically calculated time overhead for FPGA data transfer.