This project implements an image processing routine to automatically label segmented objects (blobs) in an image.
The goal is to detect, label, and individually visualize stars in the input image stars.png using a simplified Region Growing algorithm based on Flood Fill.
- Python · NumPy · Matplotlib · Scikit-Image
- Flood Fill (Region Growing)
- Binary segmentation & labeling
- Interactive visualization
- Load and convert RGB images to grayscale
- Automatically compute threshold for binarization
- Detect connected regions (stars) using
skimage.segmentation.flood() - Assign unique labels to each blob (star)
- Visualize all stars in color using
nipy_spectral - Show individual binary masks for each detected star (one per window)
- Interactive hover: Display blob ID when cursor is near a star
- Output number of detected stars in the console
Each detected star is assigned a unique label and color in the "Label Map".
Individual binary masks are shown for further analysis or visualization.
matplotlibnumpyscikit-image
Install with:
pip install matplotlib numpy scikit-imageThe script starts with this image of stars:
After grayscale conversion and thresholding, the image is binarized:
This mask (segmentation_mask.png) highlights bright areas that may contain stars.
The Flood Fill algorithm assigns a unique label to each connected region.
This is visualized using a color-coded Label Map:
Each detected star (blob) is visualized in its own grayscale binary mask.
Example – Blob 1:
This project was developed as part of a university project at FHNW and is intended for demonstration and educational purposes only.