Systems and Methods for Automated X-Ray Inspection

Simple SummaryContent extracted from patent full text and abstract with AI.

This patent describes an automated system and method for inspecting printed circuit board (PCB) assemblies using X-ray imaging combined with machine learning. The invention employs both traditional image processing and an adaptive machine learning algorithm to more accurately detect manufacturing defects—such as soldering errors—during PCB production. By capturing multiple cross-sectional X-ray images and analyzing them via a trained neural network, the system reduces false positives, minimizes manual involvement, and improves overall inspection accuracy.

Use CasesContent extracted from patent full text and abstract with AI.

• Automated quality control in electronics manufacturing lines, specifically for PCB assembly inspection.
• Detection of various soldering defects such as solder voids, bridges, or missing components in real-time during PCB production.
• Integration into end-of-line testing systems for mass production of electronic devices.
• Implementation in inspection equipment OEMs or manufacturers to enhance their product offerings.
• Application in refurbishment or repair facilities to identify and classify faults in returned or used PCBs.

BenefitsContent extracted from patent full text and abstract with AI.

• Significantly reduces manual intervention and expert setup required for inspection routines.
• Decreases false-positive rates, thus reducing unnecessary manual re-inspections and associated costs.
• Improves the speed and accuracy of defect detection compared to purely rule-based or manual systems.
• Flexible and adaptive, as the machine learning algorithm can be retrained for new defect types, PCB designs, or changes in manufacturing processes.
• Utilizes the full set of X-ray cross-sectional images, enabling more robust 3D inspection and better handling of complex PCB structures or shadowing issues.
• Scalable and suitable for high-throughput production environments, supporting continuous process optimization.

Technical Classifications (CPCs)

Inventors & Applicants

Patent Abstract

A computer-implemented method of automated X-ray inspection (10), AXI, during the production of printed circuit board, PCB, assemblies (1), comprising: Capturing (S1) an X-ray image (2) of a PCB assembly (1), Determining (S2) a first error indicator (II) based on image processing of the captured X-ray image, Determining (S3), in case the first error indicator (II) indicates the PCB assembly (1) as faulty, a second error indicator (12) based on the captured X-ray image (2) using a trained adaptive algorithm (ML), and Outputting (S4) the second error indicator (12) as a result of the inspection (10) .