Reading time ( words)
Editor’s note: This article/paper was presented at IPC APEX EXPO 2021 and was published in the Proceedings.
In recent years, neural network-based deep learning models have demonstrated high accuracy in object detection and classification in the area of digital image processing. Manufacturing industry has successfully implemented prototypes and small-scale deployment to employ artificial intelligence (AI) models for quality inspection. It has been proven that AI-assisted quality inspection can improve inspection accuracy, operation throughput, and efficiency, significantly through those prototypes and small-scale deployment. However, the industry-known challenge of Operational Technology (OT) and Information Technology (IT) integration arises when scaling up AI-assisted quality inspection in manufacturing operation. While model accuracy is the main concern from an inspection point of view, IT implementation has to meet the requirements of high availability, scalability, security, and model and device lifecycle management.
This paper discusses in detail the challenges in large-scale deployment of AI models for quality inspection operation and introduces a framework for large-scale AI-assisted quality inspection in a manufacturing environment using edge computing architecture. The framework focuses on IT architectural decisions to fulfill the OT requirement, including user experience in the quality inspection ecosystem.
Quality inspection serves as one of the critical quality assurance tools in electronics manufacturing. The execution of quality inspection is strictly governed by the control plan of the manufacturing process for a particular product. The inspection is usually performed on raw material (also known as incoming material inspection) and finished product (also known as outgoing quality inspection). In a high-complexity manufacturing process, such as wafer fabrication and integrated circuit packaging, inspection is also performed on in-process products as early in-line quality feedback. Quality inspection covers a wide range of items, from appearance, color, marking, and label, to defects and scratches.
An inspection process consists of two steps: image acquisition and image examination. Traditionally both steps are performed manually. The image is acquired directly from the product by human vision of the inspector and examined spontaneously by human cognition. As a result, traditional quality inspection is labor intensive and has great dependency on human skill and competency.
As feature sizes of interest become too small for human vision, advanced equipment such as magnifying lens, microscope, and techniques such as back lighting, dark field, and X-ray are employed to obtain the images that are examinable by human vision. Automation is enabled in image acquisition equipment and significantly improves the throughput of inspection processes for mass production. A good example is automated optical inspection (AOI) equipment used in printed circuit board inspection that has optical solution to micrometers.
With the development of computer vision, rule-based algorithms are employed to partially replace human cognition for image examination, which further improves the efficiency and throughput of the inspection process. However, rule-based algorithms have limitations in object detection and classification and are normally used as a “coarse” screen of the images under inspection. It still requires human cognition to a great extent to accurately classify the images under inspection.
In recent years, neural network-based deep learning models have demonstrated high accuracy in object detection and classification in the area of digital image processing. AI models start to show great potential to replace human cognition in the quality inspection process through object detection and classification. Thus, AI-assisted quality inspection became very promising to fully automate quality inspection processes.
To read this entire paper, which appeared in the June 2021 issue of SMT007 Magazine, click here.