一种菲涅尔透镜的外观缺陷检测方法、装置、设备和介质
By extracting grid regions from Fresnel lens images and utilizing various grayscale thresholding and morphological processing methods, the problems of uneven background brightness and separation of texture and defects in the detection of appearance defects in Fresnel lenses were solved, achieving high-precision defect identification and quality control.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- GOERTEK INC
- Filing Date
- 2023-01-09
- Publication Date
- 2026-07-17
AI Technical Summary
In existing technologies, the detection of appearance defects using Fresnel lenses suffers from problems such as uneven background imaging brightness, inconsistent hard threshold selection, and difficulty in separating textures from defects, resulting in insufficient detection accuracy.
By acquiring the grid regions suspected of having defects in the image under test, and extracting the defect contours between and overlapping textures based on the orientation of the grid regions relative to the center point of the Fresnel lens, various grayscale thresholding and morphological processing methods are used to accurately segment and identify defects.
It enables precise detection of appearance defects in Fresnel lenses, improves the accuracy of detection results, effectively identifies and intercepts defective products, and improves the quality of shipments.
Smart Images
Figure CN115931912B_ABST