# digitize-pid-yolo

**Repository Path**: dlml2/digitize-pid-yolo

## Basic Information

- **Project Name**: digitize-pid-yolo
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: main
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2026-09-09
- **Last Updated**: 2026-09-09

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# Digitize-PID Dataset (YOLO Format)

P&ID(Piping and Instrumentation Diagram) 심볼 객체 탐지를 위한 데이터셋입니다.
원본 데이터셋을 YOLO 형식에 맞게 변환하여 구성하였습니다.

## Sample

![P&ID Sample](sample.png)

## Original Source

이 데이터셋은 아래 논문과 원본 데이터를 기반으로 합니다.

> Paliwal, S., Jain, A., Sharma, M., & Vig, L. (2021). **Digitize-PID: Automatic Digitization of Piping and Instrumentation Diagrams.** *ArXiv, abs/2109.03794.*

- Paper: https://arxiv.org/abs/2109.03794
- Original Data: https://drive.google.com/drive/u/1/folders/1gMm_YKBZtXB3qUKUpI-LF1HE_MgzwfeR

## Dataset Overview

- **Images**: 500장 (train 400 / val 100, 4:1 비율)
- **Annotations**: 총 58,998개 바운딩 박스 (train 47,258 / val 11,740)
- **Format**: YOLO (`class x_center y_center width height`)
- **Classes**: 32개

## Classes

| ID | Name | ID | Name |
|----|------|----|------|
| 0 | Gate_Valve | 16 | Safety_Valve |
| 1 | Globe_Valve | 17 | Relief_Valve |
| 2 | Ball_Valve | 18 | Pressure_Relief_Valve |
| 3 | Plug_Valve | 19 | Reducer |
| 4 | Butterfly_Valve | 20 | Strainer |
| 5 | Needle_Valve | 21 | Spectacle_Blind |
| 6 | Orifice | 22 | Spacer |
| 7 | Check_Valve | 23 | Silencer |
| 8 | 3-Way_Valve | 24 | Closed_Valve |
| 9 | Angle_Valve | 25 | Indicator_General |
| 10 | Diaphragm_Valve | 26 | Rotameter |
| 11 | Double_Block_Valve | 27 | Solenoid_Valve |
| 12 | Flame_Arrestor | 28 | Differential_Indicator |
| 13 | Control_Valve | 29 | Static_Mixer |
| 14 | Butterfly_Valve_Alt | 30 | Position_Switch |
| 15 | Blind_Flange | 31 | Controller |

## Directory Structure

```
DigitizePID_Dataset/
├── data.yaml
├── images/
│   ├── train/       # 400 images
│   └── val/         # 100 images
└── labels/
    ├── train/       # 400 label files
    └── val/         # 100 label files
```

## Usage

```python
from ultralytics import YOLO

model = YOLO("yolo11n.pt")
results = model.train(data="data.yaml", epochs=100, imgsz=640)
```

## License

This dataset is provided for research purposes only. Please cite the original paper if you use this dataset.