Image processing in real-time computer vision systems using FPGA
Abstract:
One of perspective means of hardware support of collection and processing of visual data is FPGA. The greatest amount of visual data is the share of a preprocessing stage, before emergence of the first signs of quality of the image and images of objects on it. The possibility of the successful solution of tasks system of computer vision on the scale of real time significantly depends on quality and timeliness of data acquisition of preprocessing. In work results of researches on an experimental efficiency evaluation of use of FPGA as a part of means of hardware support of system of computer vision of real time are described. The general description of this means, the labor costs connected with its programming and use as a part of system of computer vision is given. The special attention is paid to the comparative analysis of time expenditure in case of implementation of algorithms of handling of images by means of the universal Intel Quad Core i7 processor of 2,5 GHz and FPGA Cyclone IV GX EP4CGX150DF31. Quantitative results of experiments on processing by the FPGA images of different objects.
