Emitter Structure Defect Classification Using Fusion Microscopy Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing emitter structures for generating X-rays are manually inspected for defects, which is labor-intensive, unreliable, and often not conducted consistently, leading to potential component failure and reduced emission efficiency.
Innovation Solution
A computer-implemented method using an artificial neural network to classify emitter structures based on fusion images from bright-field and dark-field microscopy, automatically distinguishing between defect-free and defective structures, with pixel-by-pixel analysis and noise reduction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual inspection of emitter structures is performed, then human expertise and discretion can be utilized for defect identification, but the process becomes labor-intensive, time-consuming, and unreliable
Solution Approach 1:
The patent replaces manual mechanical inspection with an automated optical inspection system that captures images of emitter structures and uses image processing algorithms to detect defects. This substitution eliminates human labor while maintaining or improving detection reliability through consistent, objective analysis of emitter images.
Solution Approach 2:
The patent creates digital copies (images) of the emitter structures and analyzes these copies to detect defects. By working with image representations rather than physical inspection, the system enables automated, high-throughput analysis without compromising detection accuracy, directly resolving the contradiction between reliability and productivity.
2Ease of manufacture
If manual inspection is performed only when defects are suspected, then inspection resources are conserved, but defect prevention and consistent quality control are compromised
Solution Approach 1:
The patent implements preliminary automated inspection of emitter structures during or after manufacturing, identifying defects before components are assembled or deployed. This preliminary detection ensures consistent quality control without disrupting manufacturing flow, as the automated system可以快速ly analyzes emitters and flags only those requiring further attention.
Solution Approach 2:
The patent establishes a feedback loop where inspection results are systematically recorded and used to monitor manufacturing quality. This continuous feedback enables consistent quality control by tracking defect patterns and alerting operators to process issues, maintaining both manufacturing efficiency and quality reliability.
3Manufacturing precision
If comprehensive manual inspection of all emitters is performed, then quality control is improved, but the labor and time requirements become unsustainable
Solution Approach 1:
The patent replaces time-consuming manual inspection with automated image capture and analysis that can process multiple emitter structures simultaneously. The system achieves high manufacturing precision through algorithmic defect detection while reducing inspection time dramatically, as computers can analyze images much faster than human operators.
Solution Approach 2:
The patent implements comprehensive automated inspection of all emitters (excessive action compared to selective manual inspection), ensuring no defects are missed. The automated system handles the full inspection load efficiently, making the previously unsustainable comprehensive approach both practical and precise.
Data Source
AI summary
At least one example embodiment relates to an apparatus, machine-readable program code, a storage medium and computer-implemented method for the automatic classification of emitter structures embodied to emit electrons for the generation of X-rays, wherein the classification takes place on the basis of a reference image of an emitter structure, wherein the classification comprises a first class and at least one second class, wherein the first class corresponds to a substantially defect-free emitter structure and the at least one second class corresponds to a defective emitter structure. Since, for the classification, an image embodied as a fusion image from a bright-field image and a dark-field image is referred to and the classification of the emitter structure into the first and the second class takes place on the basis of this image, improved defect checking of emitter structures can be provided.