通过深层嵌入和密度感知集群实现PCB缺陷的半自动标签框架
Sang-Jeong Lee1, Sung-Bal Seo2, You-Suk Bae2
1Multimodal AX Business Team, LG CNS Co., Ltd., Seoul 07795, Republic of Korea.
Sensors (Basel, Switzerland)
|October 29, 2025
概括
这项研究引入了印刷电路板 (PCB) 检查的半自动标签管道,通过有效地将异常检测建议转换为类标签,大大减少了操作员的决策.
科学领域:
- 计算机视觉 计算机视觉
- 机器学习 机器学习
- 工业自动化 工业自动化
背景情况:
- 印刷电路板 (PCB) 检查面临的挑战是由于微妙的缺陷和不平衡的数据,标签成本高,速度慢.
- 现有的方法在复杂的背景中与低对比度缺陷作斗争,阻碍了高效的质量控制.
研究的目的:
- 为PCB检查开发一个半自动标签管道,以克服数据标签瓶.
- 将异常检测输出转换为可用的类标签,提高检查开发生命周期的效率.
主要方法:
- 使用图像裁剪,可互换嵌入 (HOG,ResNet-50,ViT-B/16) 和集群 (k-means,GMM,HDBSCAN) 开发了一个管道.
- 集群级验证使用代表性质量控制组装进行.
- 评估了不同的嵌入和集群方法在缺陷分类中的有效性.
主要成果:
- 在9354个缺陷上,ResNet-50 + HDBSCAN实现了NMI ≈0.290和AMI ≈0.283的~47个集群.
- ViT-B/16 + HDBSCAN显示了可比的结果,表明了不同嵌入模型的稳定性.
- 宏观纯度超过微观纯度表明,对一次性决策进行高效的集群,可能减少运营商决策约200×.
结论:
- 拟议的工作流提供了一个可审计和灵活的路径,从异常局部化到可扩展的PCB检查监督.
- 这种方法优先考虑标签生产率,直接解决PCB检查开发中的一个关键的工业瓶.
- 半自动标签管道提高了自动视觉检查系统中缺陷分类的效率和可扩展性.
相关概念视频
Structural Classification of Joints
6.9K
Joints, also known as articulations, are classified based on their structural characteristics, i.e., based on whether the articulating surfaces of the adjacent bones are directly connected by fibrous connective tissue or cartilage, or whether the articulating surfaces contact each other within a fluid-filled joint cavity. These differences serve to divide the joints of the body into three structural classifications.
A fibrous joint is where the adjacent bones are united by fibrous connective...
A fibrous joint is where the adjacent bones are united by fibrous connective...
6.9K
Classification of Signals
1.3K
In signal processing, signals are classified based on various characteristics: continuous-time versus discrete-time, periodic versus aperiodic, analog versus digital, and causal versus noncausal. Each category highlights distinct properties crucial for understanding and manipulating signals.
A continuous-time signal holds a value at every instant in time, representing information seamlessly. In contrast, a discrete-time signal holds values only at specific moments, often denoted as x(n), where...
A continuous-time signal holds a value at every instant in time, representing information seamlessly. In contrast, a discrete-time signal holds values only at specific moments, often denoted as x(n), where...
1.3K


