LCD backlight

The ViDi™ EL Classify tool can perform not only two-type classification such as OK/NG, but also various classifications such as product type classification, sorting, and defect type classification by recognizing the characteristics of each part classification excluding individual variations. Enables higher levels of automation.

The In-Sight® 2800 vision system combines deep learning technology with traditional rule-based vision tools to solve a wide range of inspection applications. From simple presence/absence inspections to complex sorting and sorting applications, the In-Sight® 2800 makes it easy to deploy error-proof solutions. Designed for factory automation, the In-Sight® 2800 enables manufacturers of all sizes to:

Backlit lighting photography

The In-Sight® 2800 is equipped with a well-established legacy vision tools library. Many functions are available, such as distance measurement, area measurement, counting, logic operation tools, etc.

The excellent recognition power of deep learning, which enables not only OK/NG judgment but also component classification based on defect types and user-defined features, detects minute defects with high accuracy.

Back light computer

Anyway, anyone can use it easily ・The setting is completed in 3 steps. Quickly start shooting with an intuitive UI. ・No need to prepare a huge number of images. ・No GPU or programming language required. No image processing expert required. ・By switching lighting colors with multi-color lighting, it is possible to acquire the optimal image for any workpiece (optional).

Image

Backlight LED

Image

LED Backlight for TV that changes with picture

The In-Sight® 2800 is equipped with both rule-based and deep learning tools, and contributes to automation in all applications from error-proofing to inspection. In addition to using the tool alone, multiple tools can be combined to build more advanced judgment logic.

Image

The In-Sight® 2800 is equipped with a ViDi™ EL Classify tool that utilizes deep learning technology, enabling deep learning learning on the device and outputting highly accurate results in real time.

Under lighting

A modular solution of built-in illumination and lenses eliminates trial and error in choosing optical options. With just a few clicks of a button, you can decide the best configuration for your application.

Multiple inspection areas can be deployed within the field of view, and arbitrary inspection tools can be executed for each area.

Equipped with visual inspection, product type identification , and deep learning ViDi EL tools that are one rank higher, visual inspection and product type identification are possible. No need to set parameters or thresholds. ・Best class image sensor. High resolution image acquisition of 1.6 million pixels (when applicable model is selected). ・Batch inspection of multiple locations within the field of view is possible. Flexible combination of inspection tools.

The ViDi™ EL Classify tool can be used with 5-10 images per classifier and can be trained in less than 1 minute. No programming required.

The judgment reliability of the deep learning learning model can be visually confirmed in real time without any operation, helping to build highly reliable inspections.