DOI QR코드

DOI QR Code

선박안전 운항을 위한 이진 분할 알고리즘 기반 해상 객체 검출 하드웨어 가속기 설계 및 구현

Design and Implementation of a Hardware Accelerator for Marine Object Detection based on a Binary Segmentation Algorithm for Ship Safety Navigation

  • Lee, Hyo-Chan (Oceanic IT Convergence Technology Research Center, Hoseo University) ;
  • Song, Hyun-hak (Department of Information and Communication Engineering, Hoseo University) ;
  • Lee, Sung-ju (Department of Information and Communication Engineering, Hoseo University) ;
  • Jeon, Ho-seok (Department of Information and Communication Engineering, Hoseo University) ;
  • Kim, Hyo-Sung (Research Institute, San Engineering) ;
  • Im, Tae-ho (Department of Information and Communication Engineering, Hoseo University)
  • 투고 : 2020.06.30
  • 심사 : 2020.07.11
  • 발행 : 2020.10.31

초록

해상 객체 검출은 선장이 육안으로 해상 주변의 충돌 위험성이 있는 부유물을 컴퓨터를 통해 자동으로 검출하여 사람이 확인하는 방법과 유사한 정확도로 인지하는 방법을 말한다. 기존 선박에서는 레이더의 전파를 통해 해상 부유물의 유무와 거리를 판단하였지만 형체를 알아내어 장애물이 무엇인지는 판단할 수 없는 약점이 있다. 반면, 카메라는 인공지능 기술이 발달하면서 물체를 검출하거나 인식하는데 성능이 우수하여 항로에 있는 장애물을 정확하게 판단할 수 있다. 하지만, 디지털 영상을 분석하기 위해서는 컴퓨터가 대용량의 화소를 연산해야 하는데 CPU는 순차적 처리 방식에 특화된 구조이기에 처리속도가 매우 느려 원활한 서비스 지원은 물론 안전성도 보장할 수 없게 된다. 따라서 본 논문에서는 해상 객체 인식 소프트웨어를 개발하였고 연산량이 많은 부분을 가속화하기 위해 FPGA로 구현하였다. 또한, 임베디드 보드와 FPGA 인터페이스를 통해 시스템 구현 완성도를 높였으며 소프트웨어 기반의 기존 구현 방법보다 약 30배의 빠른 성능을 얻었고 전체 시스템의 속도는 약 3배 이상이 개선되었음을 확인할 수 있었다.

Object detection in maritime means that the captain detects floating objects that has a risk of colliding with the ship using the computer automatically and as accurately as human eyes. In conventional ships, the presence and distance of objects are determined through radar waves. However, it cannot identify the shape and type. In contrast, with the development of AI, cameras help accurately identify obstacles on the sea route with excellent performance in detecting or recognizing objects. The computer must calculate high-volume pixels to analyze digital images. However, the CPU is specialized for sequential processing; the processing speed is very slow, and smooth service support or security is not guaranteed. Accordingly, this study developed maritime object detection software and implemented it with FPGA to accelerate the processing of large-scale computations. Additionally, the system implementation was improved through embedded boards and FPGA interface, achieving 30 times faster performance than the existing algorithm and a three-times faster entire system.

키워드

참고문헌

  1. S. Li and K. S. Fung, "Maritime autonomous surface ships (MASS): implementation and legal issues," Maritime Business Review, vol. 4, no. 4, pp. 330-339, Nov. 2019. https://doi.org/10.1108/MABR-01-2019-0006
  2. J. S. Kang and M. S. Kang, "FPGA Implementation of ARIA Crypto-processor Based on Advanced Key Scheduling," Journal of Security Engineering, vol. 13, no. 6, pp. 439-450, Dec. 2016. https://doi.org/10.14257/jse.2016.12.05
  3. J. Lee, "Implementation and Performance Evaluation of PCI express on Xilinx FPGA," Journal of the Korea Institute of Information and Communication Engineering, vol. 22, no. 12, pp. 1667-1674, Dec. 2018. https://doi.org/10.6109/JKIICE.2018.22.12.1667
  4. T. Praczyk, "A quick algorithm for horizon line detection in marine images," Journal of Marine Science and Technology, vol. 23, no. 1, pp. 164-177, Jul. 2017. https://doi.org/10.1007/s00773-017-0464-8
  5. C. Y. Jeong, H. S. Yang, and K. Moon, "Fast horizon detection in maritime images using region-of-interest," International Journal of Distributed Sensor Networks, vol. 14, no. 7, pp. 1-11, Jul. 2018.
  6. J. R. Lee, K. R. Bae, and B. Moon, "A Hardware Architecture of Hough Transform Using an Improved Voting Scheme," The Journal of Korea Information and Communications Society, vol. 38A, no. 9, pp. 773-781, Sep. 2013.
  7. N. Otsu, "A Threshold Selection Method from Gray-Level Histograms," IEEE Transactions on Systems, Man, and Cybernetics, vol. 9, no. 1, pp. 62-66, Jan. 1979. https://doi.org/10.1109/TSMC.1979.4310076
  8. L. He, X. Ren, Q. Gao, X. Zhao, B. Yao, and Y. Chao, "The connected-component labeling problem: A review of state-of-the-art algorithms," Pattern Recognition, vol. 70, pp. 25-43, Oct. 2017. https://doi.org/10.1016/j.patcog.2017.04.018
  9. F. Bolelli, M. Cancilla, and C. Grana, "Two More Strategies to Speed Up Connected Components Labeling Algorithms," In:19th International Conference on Image Analysis and Processing, pp. 48-58, Oct. 2017.
  10. H. H. Song, H. C. Lee, S. J. Lee, H. S. Jeon, and T. H. Im, "Design of Video Pre-processing Algorithm for High-speed Processing of Maritime Object Detection System and Deep Learning based Integrated System," Journal of Internet Computing and Services, vol. 21, no. 4, pp. 117-126, Aug. 2020. https://doi.org/10.7472/JKSII.2020.21.4.117
  11. H. Rezatofighi, N. Tsoi, J. Gwak, A. Sadeghian, I. Reid, and S. Savarese, "Generalized Intersection Over Union: A Metric and a Loss for Bounding Box Regression," in 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Feb. 2019.
  12. S. Alex David, S. Ravikumar, and A. Rizwana Parveen, "Raspberry Pi in Computer Science and Engineering Education," In Intelligent Embedded Systems. Springer, Singapore, pp. 11-16, Feb. 2018.
  13. M. D. Hill and M. R. Marty, "Amdahl's Law in the Multicore Era," IEEE Computer, vol. 41, no. 7, pp. 33-38, Jul. 2008. https://doi.org/10.1109/MC.2008.209
  14. A. Cortes, I. Velez, and A. Irizar, "High level synthesis using Vivado HLS for Zynq SoC: Image processing case studies," in 2016 Conference on Design of Circuits and Integrated Systems (DCIS), Nov. 2016.
  15. K. I. Kim, "Binary Connected-component Labeling using a Multicore CPU," Journal of Korean Institute of Information Technology, vol. 11 no. 4, pp. 77-84, Apr. 2013.
  16. D. K. Prasad, D. Rajan, L. Rachmawati, E. Rajabally, and C. Quek, "Video Processing From Electro-Optical Sensors for Object Detection and Tracking in a Maritime Environment: A Survey," IEEE Transactions on Intelligent Transportation Systems, vol. 18, no. 8, pp. 1993-2016, Aug. 2017. https://doi.org/10.1109/TITS.2016.2634580