一种晶圆的缺陷检测方法及装置
By stitching together multiple views of the wafer to form a defect mosaic image and using a defect classification model combined with dynamic threshold adjustment rules, the problems of low detection accuracy and low efficiency in wafer inspection are solved, and efficient and accurate defect type identification is achieved.
CN118037704BActive Publication Date: 2026-07-17SHENZHEN ZHIXIAN FUTURE IND SOFTWARE CO LTD
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHENZHEN ZHIXIAN FUTURE IND SOFTWARE CO LTD
- Filing Date
- 2024-03-19
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
Existing technologies for wafer defect detection are affected by individual subjective factors, resulting in low detection accuracy, low efficiency, and high labor costs.
Method used
A defect mosaic is formed by stitching together defect views, magnified views, and top views of the wafer. A defect classification model is used for detection, and the defect type is determined by dynamically adjusting the threshold. The detection accuracy is improved by combining preset threshold adjustment rules.
Benefits of technology
It improves the accuracy and efficiency of wafer defect detection, reduces labor costs, and decreases the probability of classifying defects as unknown types.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN118037704B_ABST
Abstract
本说明书实施例提供一种晶圆的缺陷检测方法及装置，该方法包括：获取缺陷拼接图，其中，所述缺陷拼接图是通过拼接待检测晶圆的缺陷视图、放大视图和顶视图而得到的；将所述缺陷拼接图，输入缺陷分类模型，得到各缺陷类型对应的预测概率，其中，各缺陷类型对应的预测概率表征，缺陷拼接图存在该缺陷类型的缺陷的概率；基于各预测概率以及预设阈值调整规则，确定当前分类阈值；基于各预测概率中数值最大的第一预测概率和所述当前分类阈值，确定缺陷拼接图对应的目标缺陷类型。
Need to check novelty before this filing date? Find Prior Art