딥러닝 기반 영상 분석 알고리즘을 이용한 실시간 작업자 안전관리 시스템 개발

The purpose of this paper is to implement a deep learning-based real-time video analysis algorithm that monitors safety of workers in industrial facilities. The worker's clothes were divided into six classes according to whether workers are wearing a helmet, safety vest, and safety belt, and a...

Full description

Saved in:
Bibliographic Details
Published in스마트미디어저널 Vol. 9; no. 3; pp. 25 - 30
Main Authors 전소연(So Yeon Jeon), 박종화(Jong Hwa Park), 윤상병(Sang Byung Youn), 김영수(Young Soo Kim), 이용성(Yong Sung Lee), 전지혜(Ji Hye Jeon)
Format Journal Article
LanguageKorean
Published 한국스마트미디어학회 30.09.2020
Korean Institute of Smart Media
(사)한국스마트미디어학회
Subjects
Online AccessGet full text
ISSN2287-1322
2288-9671
DOI10.30693/smj.2020.9.3.25

Cover

More Information
Summary:The purpose of this paper is to implement a deep learning-based real-time video analysis algorithm that monitors safety of workers in industrial facilities. The worker's clothes were divided into six classes according to whether workers are wearing a helmet, safety vest, and safety belt, and a total of 5,307 images were used as learning data. The experiment was performed by comparing the mAP when weight was applied according to the number of learning iterations for 645 images, using YOLO v4. It was confirmed that the mAP was the highest with 60.13% when the number of learning iterations was 6,000, and the AP with the most test sets was the highest. In the future, we plan to improve accuracy and speed by optimizing datasets and object detection model. 본 논문에서는 산업 시설에서 작업자의 안전을 실시간으로 감시하는 딥러닝 기반 영상 분석 시스템을 구현하는 데 목적을 둔다. 작업자의 복장을 안전모, 안전조끼, 안전벨트 착용 여부에 따라 총 여섯 가지의 클래스로 나누고, 총 5,307개의 영상을 학습데이터로 이용하였다. 실험은 속도와 정확도가 준수한 YOLO v4를 이용하였으며, 총 645장의 영상에 대해 학습 반복 수에 따른 가중치를 적용했을 때의 mAP를 비교함으로써 수행되었다. 학습 반복 수 6,000에서의 mAP가 60.13%로 제일 높았으며, 테스트셋이 가장 많은 클래스의 AP가 가장 높음을 확인하였다. 추후 데이터셋과 객체 검출 모델을 최적화함으로써, 정확도와 속도를 개선할 예정이다.
Bibliography:KISTI1.1003/JNL.JAKO202028662596610
ISSN:2287-1322
2288-9671
DOI:10.30693/smj.2020.9.3.25