X-Ray Inspection Systems
An X-ray inspection sees through to the interior of the product. The possibilities of X-ray inspections are broadening to detect not only contaminants, but also product shape defects and shortages in the number of products. We are also broadening the scope of our R&D, in order to achieve greater reliability and sophistication in quality control.
Image Processing Technologies
Before signal processing, X-ray images tend to suffer from considerable noise, deriving from the shape, consistency, and other characteristics of the product. With the patient accumulation of original research, Anritsu succeeded in developing a number of powerful algorithms for signal processing and image analysis, enabling it to extract accurately just the signals generated by contaminants. Moreover, because we develop all our products in-house, we know the details of our products inside out, so we can precisely tune each component to each customer’s needs.
We offer X-ray inspection solutions that meet our customers’ needs.
X-Ray Inspection by Deep Learning
Even if feature differences are unclear in an X-ray transmission image, humans are able to detect defective products visually by comparing images of normal and defective products. However, in cases of products without constant shape and thickness, such as primary products, the patterns of conforming and defective products are extensive, and there are limits to designing algorithms based on the method of manual analysis of transmission images (making pass/fail determinations).
Deep learning processes massive amounts of pattern data, and learns and extracts features from them, enabling the implementation of algorithms that were not possible to design using conventional methods.
As a result, the judgment ranges for determining whether a product is conforming or defective can be appropriately reduced, helping to reduce inspection errors and increase yields.
