저사양 FPGA 플랫폼에서의 에지 검출 알고리즘 성능 비교 분석
Performance Comparison and Analysis of Edge Detection Algorithms on Low-End FPGA Platforms
2
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
Edge detection is a crucial image-processing technique used in autonomous driving, medical imaging, and military applications, including weapon systems for target tracking and detection. This study analyzed the real-time implementation of edge detection algorithms—Roberts, Sobel, Laplacian, and Laplacian of Gaussians (LoG)—on embedded systems using Vitis HLS on an FPGA platform. We compared these algorithms based on latency, resource utilization, and power consumption and evaluated their performance at HD (720p) and FHD (1080p) resolutions. The results indicated that the Roberts filter had the lowest latency, whereas the Laplacian filter was the most efficient in terms of resource usage and power consumption. The Sobel filter provided the best edge detection performance but used more resources. The LoG filter was robust against noise but had the highest latency and resource usage owing to its Gaussian operation. The HD filters had a latency approximately 2.2 times shorter than that of the FHD filters, with minimal differences in resource usage and power consumption. These findings provide essential guidance for selecting edge detection algorithms suitable for real-time embedded applications, balancing efficiency and performance.
Abstract
Edge detection is a crucial image-processing technique used in autonomous driving, medical imaging, and military applications, including weapon systems for target tracking and detection. This study analyzed the real-time implementation of edge detection algorithms—Roberts, Sobel, Laplacian, and Laplacian of Gaussians (LoG)—on embedded systems using Vitis HLS on an FPGA platform. We compared these algorithms based on latency, resource utilization, and power consumption and evaluated their performance at HD (720p) and FHD (1080p) resolutions. The results indicated that the Roberts filter had the lowest latency, whereas the Laplacian filter was the most efficient in terms of resource usage and power consumption. The Sobel filter provided the best edge detection performance but used more resources. The LoG filter was robust against noise but had the highest latency and resource usage owing to its Gaussian operation. The HD filters had a latency approximately 2.2 times shorter than that of the FHD filters, with minimal differences in resource usage and power consumption. These findings provide essential guidance for selecting edge detection algorithms suitable for real-time embedded applications, balancing efficiency and performance.
참고문헌 (14)
- Hasan, M. H., "Analysis of Edge Detection for Road Lanes through Hardware Implementation", Proc. IEEE Int. Conf. Electr. Comput. Commun. Technol. (ICECCT), pp.1~6 (2021)
- Quoc Thai, P., "Application of Edge Detection Algorithm for Self-Driving Vehicles", Proc. Natl. Sci. Conf. Appl. New Technol. Green Build (ATiGB), pp.225~228 (2022)
- Shao, Y., "Edge Detection Algorithm of MRI Medical Image Based on Artificial Neural Network", Proc. Comput. Sci., 208, pp.136~144 (2022)
- Roberts, L. G., "Machine Perception of Three-Dimensional Solids", Massachusetts Institute of Technology, 1~82 (1963)
- Sobel, I., "A 3x3 Isotropic Gradient Operator for Image Processing", A talk at the Stanford Artificial Project, 271~272 (1968)
- Prewitt, J. M., "Object Enhancement and Extraction in Picture Processing and Psychopictorics", pp.75~149 (1970), Academic Press
- Marr, D., "Theory of Edge Detection", Proc. R. Soc. Lond. B Biol. Sci, 207(1167), pp.187~217 (1980)
- Xie, S., "Holistically-Nested Edge Detection", Proc. IEEE Int. Conf. Comput. Vis. (ICCV), pp.1395~1403 (2015)
- LeCun, Y., "Deep Learning", Nature, 521, pp.436~444 (2015)
- Simonyan, K., "Very Deep Convolutional Networks for Large-Scale Image Recognition", Proc. Int. conf. Learn. Represent.(ICLR), pp.1~24 (2015)
- Canny, J., "A Computational Approach To Edge Detection", IEEE Transactions on Pattern Analysis and Machine Intelligence, 8(6), pp.679~698 (1986)
- W. Rong, "An improved Canny edge detection algorithm", Proc. IEEE Int. Conf. Mechatronics and Automation (ICMA), pp.577~582 (2014)
- Pratt, W. K., "Digital Image Processing" (1978), John Wiley &Sons Inc
- Gonzalez, R. C., "Digital Image Processing" (2002), Prentice Hall