Via Hole Defect Detection Using Image Clustering and Edge Expansion
Solution Overview
Problem
Residues in etched via holes of semiconductor devices can reduce conductivity, and existing methods lack an efficient automated detection process to ensure 100% inspection without human intervention.
Innovation Solution
An automatic detecting apparatus and method that performs image processing steps, including clustering, contouring, boundary analysis, edge detection, and expansion, to identify residues in via holes from top-view images, enabling 100% inspection without human power.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual inspection methods are used to detect residues in via holes, then detection accuracy can be maintained, but inspection efficiency and productivity are greatly reduced
Solution Approach 1:
The patent replaces manual mechanical inspection with an automated image processing system that uses clustering algorithms, contour detection, and edge analysis to automatically identify residues in via holes, thereby eliminating human labor while maintaining detection capability
Solution Approach 2:
The patent creates a digital copy of the via hole structure through image processing, generating a binary image and defect pattern that replicates the physical structure, allowing automated analysis without direct physical inspection
2Productivity
If automated detection methods are implemented, then inspection productivity is improved, but measurement precision and detection accuracy may deteriorate
Solution Approach 1:
The patent segments the via hole image into distinct regions using clustering algorithms that divide pixels into different clusters based on intensity values, allowing precise identification of residue regions while maintaining high processing speed
Solution Approach 2:
The patent applies different processing strategies to different regions of the image, using contour detection for the via hole boundary and edge detection for residue identification, optimizing detection accuracy for each specific feature
3Measurement precision
If complex image processing steps are applied to improve detection accuracy, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent divides the image processing task into distinct sequential steps (clustering, contouring, edge detection, pattern generation), where each step handles a specific aspect of the detection problem, making the overall complex system manageable and implementable
Solution Approach 2:
The patent introduces intermediate data structures such as the binary image and defect pattern as mediators between the raw image input and the final detection result, simplifying the processing logic at each stage
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
An automatic detecting method and an automatic detecting apparatus using the same are provided. The automatic detecting apparatus includes an inputting unit, a dividing unit, a contouring unit, a range analyzing unit, a boundary analyzing unit, an edge detecting unit, an expanding unit and an overlapping unit. The dividing unit is used for dividing an overlooking image into four clusters via a clustering algorithm. The contouring unit is used for obtaining a contour. The range analyzing unit is used for obtaining a detecting range. The boundary analyzing unit is used for obtaining a circular boundary in the detecting range. The edge detecting unit is used for obtaining a plurality of edges in the circular boundary. The expanding unit is used for expanding the edges to obtain a plurality of expanded edges. The overlapping unit is used for overlapping the expanded edges and the contour to obtain a defect pattern.