| 张 恒,程 成,袁 彪,赵洪坪,吕 雪,杭 芹.基于 YOLOv5-EA-FPNs 的芯片缺陷检测方法研究[J].电子测量与仪器学报,2023,37(5):36-45 |
| 基于 YOLOv5-EA-FPNs 的芯片缺陷检测方法研究 |
| Research on chip defect detection method based on YOLOv5-EA-FPNs |
| |
| DOI： |
| 中文关键词: 芯片缺陷检测 深度学习 特征金字塔 多尺度融合 小目标检测 YOLOv5 |
| 英文关键词:chip defect detection deep learning FPNs multi-scale fusion small object detection YOLOv5 |
| 基金项目:国 家 自 然 科 学 基 金 项 目 ( 12005030)、 重 庆 市 自 然 科 学 基 金 ( cstc2021jcyj-bsh0252)、 磁 约 束 聚 变 安 徽 省 实 验 室 开 放 基 金
(2021AMF01004)项目资助 |
|
|
|
|
| 摘要点击次数: 396 |
| 全文下载次数: 692 |
| 中文摘要: |
| 针对芯片缺陷检测中,缺陷尺寸跨度大、特征相似、小目标难识别、漏检等问题,本文提出基于 YOLOv5 改进的缺陷检测
方法。 针对小目标缺陷检测中出现的漏检、误检等问题,提出新增小目标特征检测器(small target feature detector, S-Detector),
提升模型对小目标缺陷的学习能力;针对缺陷尺寸跨度大、特征相似等问题,提出具有高效聚焦学习能力的特征金字塔结构
(efficient attention feature pyramid networks, EA-FPNs),提升模型对不同尺寸缺陷的检测能力;针对预测阶段冗余框较多导致时
间开销大的问题,提出基于面积的边界框融合算法(bounding box fusion algorithm, BFA),减少冗余框。 实验结果表明,本文方法
相较于改进前,检测精确度提升 1. 2%,小目标缺陷精确度提升 1. 6%;采用 BFA 消除冗余框的同时,平均检测时长为 26. 8 μs/
张,较使用 BFA 前减少了 5. 2 μs。 本文所提方法具有良好性能,能够提升检测效率。 |
| 英文摘要: |
| To address the problems of large defect size span, similar characteristics, difficulty in recognition of small targets, and missed
objects in chip defect detection, an improved method based on YOLOv5 is proposed. To solve missed and false detection of small
targets, we presented a new small target feature detector ( S-Detector) to improve the learning capability of the model. For the large
defect size span and similar characteristics, efficient attention feature pyramid networks (EA-FPNs) with highly active focus learning
ability are proposed to improve the ability to detect different sizes of defects. The bounding box fusion algorithm (BFA) is developed to
reduce the redundant boxes and time overhead in prediction. The experimental results show that the detection accuracy of this method is
enhanced by 1. 2% and the accuracy of minor target defects is improved by 1. 6%; while using BFA to eliminate the redundant boxes,
the detection time of a single image is 26. 8 μs, which is decreased by 5. 2 μs before BFA. The proposed method has good performance
and efficiency in chip defect detection.
Keywords:chip defect detection; deep l |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
