PCB Assembly Inspection Using Machine Learning for Early Defect Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing complexity and variety of printed circuit board (PCB) assemblies lead to errors during component placement, resulting in defective PCBs due to workers confusing layouts, which are often discovered only after soldering, causing material waste and inefficiency.
Innovation Solution
Implementing an automated optical inspection system that uses a camera to acquire images of PCB assemblies before soldering, employing object-based analysis and machine learning algorithms to recognize components and compare them against stored assembly information, allowing for real-time correction of placement errors and preventing defective PCBs from progressing to the soldering stage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual inspection is performed after soldering, then defects can be detected, but material waste and production loss increase
Solution Approach 1:
The patent implements inspection before soldering by capturing images of the PCB assembly at the placement stage. The image processing system analyzes component placement accuracy, polarity, and presence before the soldering process occurs. This preliminary inspection allows defective assemblies to be identified and corrected before material commitment through soldering, thereby reducing material waste while maintaining defect detection capability
Solution Approach 2:
The patent introduces an intermediary image processing system between the placement and soldering processes. This system captures images, processes them through object-based analysis and machine learning algorithms, and provides feedback on placement quality. The intermediary system enables early defect detection without requiring manual inspection after soldering, thus preventing material waste while ensuring reliability
2Adaptability or versatility
If high variety of PCB assemblies are produced, then product versatility increases, but worker errors and defects increase
Solution Approach 1:
The patent implements a self-service inspection system where the image processing algorithm automatically identifies and verifies component placement on PCB assemblies. The system uses object-based analysis and machine learning to autonomously detect placement errors, polarity issues, and missing components without human intervention. This automated self-inspection ensures consistent quality across high-variety PCB productions, eliminating worker errors while maintaining adaptability to different assembly types
Solution Approach 2:
The patent establishes a feedback loop where the image processing system continuously monitors placement quality and provides real-time information about defects. The system compares detected components against the bill of materials and placement coordinates, generating feedback on placement accuracy. This automated feedback mechanism ensures high reliability across diverse PCB assemblies by immediately identifying and flagging placement errors regardless of assembly variety
3Reliability
If inspection is performed after soldering, then defect detection is possible, but production efficiency decreases
Solution Approach 1:
The patent performs inspection at the placement stage before soldering, capturing images and analyzing component placement early in the production process. This preliminary action allows defects to be detected and corrected before the time-consuming soldering operation, thereby maintaining defect detection capability while significantly improving production efficiency by preventing rework after soldering
4Loss of substance
If automated optical inspection is implemented before soldering, then material utilization improves, but system complexity increases
Solution Approach 1:
The patent replaces complex manual inspection processes with an automated optical inspection system using machine learning algorithms. The system captures images with a camera and uses software-based object recognition to analyze component placement. This substitution of mechanical/manual inspection with automated optical-mechanical systems improves material utilization through early defect detection while the software-based approach keeps hardware complexity manageable
Data Source
AI summary
A method of inspecting a printed circuit board (PCB) assembly includes acquiring an image of the PCB assembly and analyzing the image, wherein the analysis includes an object-based analysis of the image for recognizing at least one component placed on the PCB, wherein the object-based analysis is performed based on an object-based analysis program, and wherein the object-based analysis program includes a trained machine learning model. The method further includes: determining whether the at least one component is placed on the PCB based on a comparison between a finding of the object-based analysis and stored assembly information for the PCB; outputting an error when one or more components are missing or wrongly placed; inputting a result of a visual inspection of the PCB assembly that indicates a pseudo-error of the object-detection analysis; and writing one or more settings for soldering the PCB assembly by a soldering device.