Scraped surface quality visual detection method based on distance field high point identification
By employing a visual inspection method based on high-point recognition in the distance field, utilizing Lab color space and Euclidean distance field technology, the problems of insufficient human experience and environmental interference in scraping inspection are solved, achieving accurate recognition of high points and improving lubrication properties.
Patent Information
- Application Number
- CN202610541145.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-04-22
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies rely on human experience in scraping and testing, resulting in a lack of objective consistency in test results and difficulty in effectively identifying high points of large-area adhesion and high points under environmental interference.
A visual detection method based on distance field high point recognition is adopted. The brightness and blue-yellow components are extracted by transforming the Lab color space, a composite scoring function model is constructed, and the high point region is divided by combining the Euclidean distance field and adaptive dynamic threshold segmentation. High points are then filtered by roundness and aspect ratio to achieve accurate high point recognition.
It improves the accuracy and consistency of high point identification, reduces manual intervention, effectively separates adhered high points, and improves detection efficiency and lubrication properties.
Smart Images
Figure CN122415527A_ABST