A wafer frame detection method, device, equipment and storage medium
By identifying target wafers and defects in wafer images, zeroing the grayscale value of defect areas, and adjusting the target bounding box by combining overall and local grayscale values, the detection accuracy problem of existing models under extreme conditions is solved, achieving higher precision and adaptive wafer bounding box detection.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- STORAGEX TECH INC
- Filing Date
- 2024-03-05
- Publication Date
- 2026-07-21
AI Technical Summary
Existing deep learning-based wafer edge detection models are prone to false positives or false negatives when there is a lack of large amounts of labeled data, extreme lighting, or occlusion. They also have high computational resource requirements, poor model interpretability, and limited generalization ability to unknown patterns, resulting in low detection accuracy.
By identifying the target wafer and defects in the wafer image, zeroing the grayscale value of the defect area, calculating the average grayscale value of the target wafer and the border, and adjusting the target box based on the grayscale value, a modified box is generated. Taking into account overall and local grayscale information, the target box is dynamically adjusted to improve detection accuracy.
It improves the accuracy and adaptability of wafer edge detection, enhances adaptability to different regions, reduces interference from defects in detection, and improves the accuracy and stability of edge recognition.
Smart Images

Figure CN117893835B_ABST