Adaptive X-ray Inspection for PCB Assembly Defects
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current automated X-ray inspection systems for electronics production require significant manual effort and expertise for setting up inspection routines, leading to high false positive rates and inefficiencies.
Innovation Solution
A computer-implemented method for automated X-ray inspection that captures X-ray images of PCB assemblies, processes them using conventional image processing to determine a first error indicator, and then uses a trained adaptive algorithm to determine a second error indicator, reducing the need for manual intervention and improving accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual setup routines are used for X-ray inspection, then inspection accuracy can be maintained through expert knowledge, but significant manual effort and time are required for setup and adjustment
Solution Approach 1:
The system performs preliminary actions by automatically capturing X-ray images and generating 3D models of the PCB assembly before inspection. The adaptive algorithm is pre-trained on historical inspection data, enabling it to automatically evaluate solder joints without requiring manual setup of inspection routines for each new product type.
Solution Approach 2:
The inspection system serves itself by using the trained adaptive algorithm to automatically evaluate inspection results without human intervention. The system self-adjusts by learning from false positives and negatives, continuously improving its accuracy without requiring manual reconfiguration by experts.
2Stability of the object's composition
If deterministic rule-based systems are used for inspection evaluation, then consistent evaluation criteria are applied, but high false positive rates occur due to inability to capture complex defect patterns
Solution Approach 1:
The system transitions from fixed deterministic rules to adaptive parameters that can dynamically adjust based on learned patterns. The adaptive algorithm modifies evaluation criteria by learning from training data, allowing it to capture complex defect patterns while maintaining consistent evaluation through the learned model.
Solution Approach 2:
The evaluation system combines multiple approaches: traditional deterministic rules provide a baseline for consistent evaluation, while the trained adaptive algorithm (using machine learning models) adds the capability to recognize complex patterns. This composite approach maintains stability while reducing false positives.
3Productivity
If only one layer of grayscale image is inspected, then processing time is reduced, but valuable information is lost
Solution Approach 1:
The system extracts the most informative features from the 3D model and multi-layer grayscale images, selecting only the critical information needed for inspection. This allows processing of comprehensive 3D data without the full processing time penalty, extracting essential solder joint information while discarding redundant data.
Solution Approach 2:
The system moves from inspecting only a single 2D grayscale image layer to utilizing 3D models generated from multiple X-ray images. This adds the depth dimension, allowing inspection of solder joints from multiple angles and planes, capturing information that would be invisible in a single 2D layer while using efficient rendering to maintain processing speed.
4Measurement precision
If separate 3D processes are used for PCB inspection, then shadowing effects are reduced, but system complexity and processing time increase
Solution Approach 1:
The system merges the 3D modeling process with the inspection evaluation process. Instead of treating them as separate operations, the 3D model generated from X-ray images is directly integrated into the adaptive algorithm's evaluation workflow, allowing the same system to perform both functions without requiring separate complex 3D inspection equipment.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The proposed solution reduces the number of false positives during PCB assembly inspection, decreases the setup efforts required for inspection, and allows for more efficient use of X-ray images by utilizing the full sequence of cross-sectional images.
Implementation Method 1
Due to radiation as it passes through matter, a gray scale image (0-255) may be generated by a digital sensor
Data Source
AI summary
A computer-implemented method of automated X-ray inspection during the production of printed circuit board, PCB, assemblies. The method includes capturing an X-ray image of a PCB assembly, determining a first error indicator based on image processing of the captured X-ray image, determining, in case the first error indicator indicates the PCB assembly as faulty, a second error indicator based on the captured X-ray image using a trained adaptive algorithm, and outputting the second error indicator as a result of the inspection.