Earticle

초록

영어

Currently, power plants are using conveyor belts that are used to transport large amounts of coal. However, there is a risk of fire due to the falling coal that occurs during transport. Currently, human resources are deployed to prevent accidents through patrol inspections, but due to the wide range of transport facilities, there is no way to check the status of droppings occurring in all places. Therefore, in this paper, we proposed a method that can always monitor the height of a drop and provide it by determining the height by using a CNN algorithm that shows high accuracy and performance in recent image analysis. A 1-Stage object detection neural network that can be used for the detection of a drop was selected, and basic performance verification of the study was performed using the inference model generated through retraining. The homework on how to more accurately calculate the area of the drop remains, but it is expected to contribute to automating the monitoring of the drop over a wide range of areas that were performed by putting manpower into the field in the future.

목차

Abstract
I. 서론
II. 관련 연구
1. 배경 제거
2. 영상처리 기반 객체 검출
3. CNN 기반 객체 검출
III. 본론
Ⅳ. 구현 및 실험 결과
V. 결론
[ 참고문헌 ]

키워드

저자

• 조익현 [ Ik-Hyun Jo | ㈜싸인텔레콤 부설 연구소 ] 교신저자
• 박영기 [ Young-Ki Park | ㈜싸인텔레콤 대표이사 ]

참고문헌

발행기관

간행물

표지보기 관심저널 등록

• 간행물명

  정보통신설비학회논문지 [Journal of the Institute of Information and Telecommunication Facility Engineering]

• 간기

  계간

• pISSN

  1598-6586

• 수록기간

  2002~2020

• 십진분류

  KDC 567 DDC 621