P&ID Line Detection Dataset

This dataset contains cropped images from P&ID (Piping and Instrumentation Diagrams) with line segment annotations for line detection and segmentation tasks.

Dataset Description

• Total source images: 500
• Total cropped samples: 10000
• Total line segments: 44754
• Crops per image: 20
• Image sizes: Various sizes under 1000px (e.g., 300x500, 512x768, 900x900)

Dataset Structure

Each sample contains:

• file_name: Image filename
• source_image_idx: Index of the original P&ID image
• crop_idx: Index of this crop from the source image
• width: Crop width in pixels
• height: Crop height in pixels
• lines: Dictionary with:
  • segments: List of line segments as [x1, y1, x2, y2] (start and end points)
  • line_types: List of line types ("solid" or "dashed")
  • pipelines: List of pipeline names for each line

Usage

from datasets import load_dataset

# Load the dataset
dataset = load_dataset("imagefolder", data_dir="path/to/lines_dataset")

# Access a sample
sample = dataset["train"][0]
image = sample["image"]
lines = sample["lines"]
segments = lines["segments"]      # [[x1, y1, x2, y2], ...]
line_types = lines["line_types"]  # ["solid", "dashed", ...]
pipelines = lines["pipelines"]    # ["5\"-EK-2648", ...]

# Draw lines on image
from PIL import ImageDraw
draw = ImageDraw.Draw(image)
for seg in segments:
    x1, y1, x2, y2 = seg
    draw.line([(x1, y1), (x2, y2)], fill="blue", width=3)
image.show()

Line Segment Format

Each line segment is represented as [x1, y1, x2, y2] where:

• (x1, y1) is the start point
• (x2, y2) is the end point
• Coordinates are in pixels, relative to the cropped image
• Only lines where both endpoints are fully inside the crop are included

Line Types

• solid: Continuous pipeline lines
• dashed: Dashed lines (often representing signal/instrument lines)

License

MIT License