Circuit Board Inspection via Near-Infrared Imaging and Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current inspection methods for circuit board assemblies are inadequate for quickly and efficiently detecting anomalies, especially in internal metal interconnects, and often require time-consuming x-ray tomography or are limited to surface-level visible light inspections.
Innovation Solution
The use of near-infrared (NIR) imaging combined with machine learning models to analyze NIR images of circuit boards, enabling the detection of anomalous features and reducing the need for x-ray imaging by capturing internal metal interconnects and chip components not visible in visible light.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If visible light inspection is used, then inspection speed is fast and large number of boards can be inspected, but internal metal interconnects and chip components cannot be detected
Solution Approach 1:
The patent introduces near-infrared radiation as an intermediary medium to penetrate the circuit board and reveal internal metal interconnects. The NIR imaging system acts as a mediator between the inspector and the hidden internal features, allowing detection without physical contact or disassembly of the board.
Solution Approach 2:
The patent replaces the mechanical/x-ray based inspection methods with optical near-infrared imaging. Instead of using heavy x-ray equipment or physical sectioning of boards, the system uses NIR light to achieve internal feature detection, substituting a simpler optical system for complex mechanical or radiological systems.
2Measurement precision
If x-ray tomography is used, then all layers and solder joints can be imaged, but inspection is time-consuming and limited to few samples
Solution Approach 1:
The patent substitutes x-ray tomography with near-infrared imaging to achieve internal feature detection. The NIR system captures images in real-time or near real-time, eliminating the time-consuming nature of x-ray tomography while maintaining the capability to image internal layers and solder joints.
Solution Approach 2:
The patent creates optical copies (NIR images) of the internal circuit board structure without requiring physical sectioning or complex tomographic reconstruction. The NIR imaging system produces direct visual copies of internal features, replacing the need for time-intensive x-ray tomographic processing.
3Measurement precision
If x-ray imaging is used, then internal features can be detected, but cost and acquisition time are high
Solution Approach 1:
The patent employs inexpensive near-infrared imaging equipment instead of costly x-ray imaging systems. The NIR cameras and light sources are significantly cheaper than x-ray equipment, making the inspection process more cost-effective while maintaining the ability to detect internal features.
Solution Approach 2:
The patent replaces expensive x-ray imaging infrastructure with affordable near-infrared optical systems. This substitution dramatically reduces both equipment costs and operational expenses while achieving the same functional goal of internal feature detection.
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
This approach allows for faster and more cost-effective inspection of circuit boards, enhancing the detection of assembly errors, malicious components, and intellectual property infringement by visualizing internal features with high contrast and accuracy.
Implementation Method 1
shining a near-infrared (NIR) light on the board and capturing an image of the board with a camera
Data Source
AI summary
Circuit board inspection by receiving a near infrared (NIR) image of at least a portion of a circuit board, analyzing the NIR image using a machine learning model, and detecting anomalous circuit board portions according to the analysis.