• Title/Summary/Keyword: Undersea image

Search Result 3, Processing Time 0.018 seconds

Semantic Segmentation of the Habitats of Ecklonia Cava and Sargassum in Undersea Images Using HRNet-OCR and Swin-L Models (HRNet-OCR과 Swin-L 모델을 이용한 조식동물 서식지 수중영상의 의미론적 분할)

  • Kim, Hyungwoo;Jang, Seonwoong;Bak, Suho;Gong, Shinwoo;Kwak, Jiwoo;Kim, Jinsoo;Lee, Yangwon
    • Korean Journal of Remote Sensing
    • /
    • v.38 no.5_3
    • /
    • pp.913-924
    • /
    • 2022
  • In this paper, we presented a database construction of undersea images for the Habitats of Ecklonia cava and Sargassum and conducted an experiment for semantic segmentation using state-of-the-art (SOTA) models such as High Resolution Network-Object Contextual Representation (HRNet-OCR) and Shifted Windows-L (Swin-L). The result showed that our segmentation models were superior to the existing experiments in terms of the 29% increased mean intersection over union (mIOU). Swin-L model produced better performance for every class. In particular, the information of the Ecklonia cava class that had small data were also appropriately extracted by Swin-L model. Target objects and the backgrounds were well distinguished owing to the Transformer backbone better than the legacy models. A bigger database under construction will ensure more accuracy improvement and can be utilized as deep learning database for undersea images.

A Study on Application Method of Parallel Processing for Performance Improvement of Sonar-based Undersea Simulation (소나 기반 해저 시뮬레이션의 성능 향상을 위한 병렬처리 적용 방법 연구)

  • Back, Seoung-Jea;Lee, Keon-Pyo;Ha, Ok-Kyoon
    • Proceedings of the Korean Society of Computer Information Conference
    • /
    • 2018.07a
    • /
    • pp.1-2
    • /
    • 2018
  • 해상 선박의 안전을 위해 해저의 객체 및 장애물의 정확한 탐지를 위해 해저환경에서 감쇠현상이 비교적 적은 음파 기반의 소나가 널리 활용된다. 그러나 기존의 소나 영상 시뮬레이션은 고해상도의 영상, 잡음 처리, 해저지형과 객체 데이터 등의 방대한 데이터 처리로 인해 물체 탐지 및 식별을 위한 처리속도와 비용이 크게 증가한다. 이러한 문제를 최소화하기 위해서 해저지형, 객체 생성과 잡음 처리 모델을 Multi-Threading, SIMD 등 병렬처리를 적용하여 처리속도를 최적화 한다. 본 논문에서는 혼합된 병렬처리 방법을 적용하여 소나를 기반으로 해저 환경 시뮬레이션을 위한 모의 신호를 생성하는 성능을 향상시킨다. 병렬처리로 인해 개선된 성능을 순차처리에 따른 속도와 실험적으로 비교한다.

  • PDF

Development of Smart-phone based Thermal Imaging Diagnostic System for Monitoring Disc Pads of Crane (크레인 디스크 패드 모니터링을 위한 스마트폰 기반의 열영상 진단 시스템 개발)

  • Oh, Yeon-Jae;Park, Kyoung-Wook;Kim, Eung-Kon
    • The Journal of the Korea institute of electronic communication sciences
    • /
    • v.9 no.12
    • /
    • pp.1397-1404
    • /
    • 2014
  • Grab cranes are used for multi-purpose when the sand and soil are deposited into harbor wharf or the undersea construction is performed. Among the components of crane grab, the wire drum and disc brake pad are key expendables and have disadvantages that lot of heat is generated and very expensive when replacing them. In this study, the thermal image analysis for the disc brake, which works with wire drum of the crane is suggested. The suggested system performs the pad thermal diagnosis through the thermal image using the characteristics that the disc and pad surface temperatures are distributed abnormally before the brake failure and the disc pad damage. Therefore, the damage by the failure can be prevented by discovering the abnormality of the machine parts before failure and the life cycle of the pad and the cost can be extended and saved by operating the crane performing constant checkup for the overload.