| 학회 | 한국공업화학회 |
| 학술대회 | 2020년 가을 (10/28 ~ 10/30, 광주 김대중컨벤션센터(Kimdaejung Convention Center)) |
| 권호 | 24권 1호 |
| 발표분야 | 포스터-화학공정 |
| 제목 | Deep learning for an image recognition of plant engineering diagram |
| 초록 | A piping and instrumentation diagram (P&ID) is a core drawing in the EPC(Engineering, Procurement, Construction) industry because it contains information about the units and instrumentation of the plant. Until now, simple repetitive tasks, such as listing symbols in P&ID drawings, have been done manually. Currently, a deep learning model based on CNN(Convolutional Neural Network) has been studied for drawing object detection, but the detection time is too long and the accuracy is not satisfactory to be implemented in the practical application. In this study, the detection of symbols in a drawing was performed using the 1-stage object detection algorithm, YOLO(You Only Look Once). Specifically, we built the training data using the image labeling tool, and trained it on the model. Next, the performance of the model was compared with the conventional recognition algorithms. Acknowledgement : This research was supported by the Seoul R&BD Program(CY190032). |
| 저자 | 신호진, 전은미, 이철진 |
| 소속 | 중앙대 |
| 키워드 | Piping and Instrumentation Diagram; Object Detection; Plant Engineering; Deep Learning |
