Glossary · Machine vision
Blob analysis
Also known as: Connected component analysis, Region analysis
German: Blob-Analyse
In image processing, blob analysis identifies connected regions of pixels (blobs) that share a property, usually after thresholding, and measures features such as area, position, perimeter or shape for each region.
- Machine vision
In one sentence
Blob analysis finds connected pixel regions after thresholding and measures their area, position and shape for counting and checks.
Example
Blob analysis counts the tablets in each blister pocket image and flags pockets whose blob area is too small, indicating a broken tablet.
How it applies
- Engineering: Blob analysis is fast and robust when contrast is good, for example with Backlighting. It is used for counting, sorting, presence checks and simple defect detection. Features such as area, circularity or bounding box then serve as pass/fail criteria for each region.
- Commissioning: Threshold, minimum blob size and feature limits are tuned with good and bad sample parts; lighting changes can shift the results.
- Documentation: The documentation team should record thresholds and limits with the sample parts used to set them, and describe which parameters operators may adjust.
Blob analysis vs. pattern matching
Blob analysis evaluates regions without a model and suits simple, high-contrast scenes. Pattern matching searches for a trained model and handles cluttered backgrounds and rotated parts better.