Going Beyond Pass/Fail with Automatic Defect Recognition (ADR)
June 23, 2016
X-ray inspection has long been a preferred way to obtain non-destructive quality data on ...
X-ray inspection has long been a preferred way to obtain non-destructive quality data on manufactured parts. With the advent of digital X-ray systems, improved imaging capabilities, and more powerful computers, many manufacturers have come to rely on non-destructive X-ray test systems to make pass/fail decisions on individual parts. This is called Automatic Defect Recognition or ADR.
Although previous YXLON publications have described the general principles of ADR, our latest paper, "Going Beyond Pass/Fail with Automatic Defect Recognition" focuses on two particular applications of ADR technology: fully automated X-ray inspection as part of a manufacturing process and ensuring increased security via ADR.
This paper examines the considerations related to implementing ADR for production line inspection, with particular focus on metal casting inspection and automotive applications. We outline six phases of ADR implementation that exceed a simple pass or fail determination, which is typical in "one-off" inspection routines:
- Manipulation and Infrastructure
- Image Acquisition
- Image Processing
- Decision Optimization
- Process Optimization
- Learning and Adaptation
ADR is capable of delivering benefits that impact business profitability. Specifically, ADR for manufacturing can help reduce scrap or wasted parts, increase speed of inspection, improve manufacturing processes, and enhance the security of parts that are produced.
In the context of ADR, "security" refers to the improved confidence operators can achieve about the accuracy of pass/fail decisions. This is relevant for any industrial application where releasing faulty parts into the market risks harming people or property, such as aerospace applications, consumer/end user products and automotive supply chain manufacturing.
What is ADR?
Defects that were difficult to detect when using older techniques are now much easier to identify with Automatic Defect Recognition (ADR). With the appropriate software tools, systems can help inspectors find and characterize anomalies, and even automatically accept or reject the part based on the system settings. Automatic or Automated Defect Recognition is defined as:
- Automatic: No operator is required: fully automatic inspection with fully automatic decision-making.
- Assisted: The system processes the images and indicates potential defects to the operator, who then uses this information to make the final decision.
- Defect: The system is taught or can infer what constitutes a defect in a specific application.
- Recognition: The software analyzes the X-ray image and makes a decision or recommendation using characterizations, measurements, etc
More Than Just Image Processing
ADR involves more than just sophisticated processing of images. If the goal of the metal casting operation or any other type of manufacturing business is to leverage ADR fully, managers should approach ADR as a holistic process. ADR is more than just image processing:
When successfully applied, ADR enhances efficiency at every stage of the cycle shown above. From manipulation and infrastructure to greater learning and adaptation, ADR gives managers and operators the tools they need to make intelligent adjustments. These critical adaptations improve their learning, reduce production cycle time, improve quality, and contribute to statistical process control (SPC) programs. Continual repetition of the process also helps increase return on investment (ROI) from the ADR investment.
ADR should not be considered solely as software that augments a machine or system. In fact, ADR can be the heart of a complete inspection process. The various stages that make up this concept are described in this publication.
Read the full paper to learn more about ADR and how it can benefit your inspection efficiency and profitabiltiy. Read the comprehensive white paper on this topic, a free PDF download.
June 02, 2022 | Jeff Urbanski, Peter Koch
Electronics manufacturing companies face the challenge of precise porosity evaluation inside their products. Peter Koch and Jeff Urbanski show new analysis methods working together for more accurate void measurement at critical interfaces.Read more