Deep Learning
Realizes truly optimized image inspection at our customers’ sites.
ANY Deep Learning Package.
Characteristics of deep learning at ANY
・Image inspection closer to that of human eyes (sensory evaluation)・Automation of previously difficult processes to save on the usage of personnel resources
・Intuitive operation with settings that can be changed by the customer
Finally, a tool that can replace visual inspections.
“Lastly, we need to use human eyes.”
If your company has been disillusioned with image inspection up until now, deep learning provided by ANY is a tool that will bring about revolutionary changes for your company.
In conventional rule-based image inspection, for uses with variation in error decision and where identification with small differences in accuracy is unfeasible, enormous case studies must be input/created. This is a factor leading to long development periods and higher costs. There are other issues, such as changes in the specifications of the product or improvements to the production method necessitating reprogramming, causing the merits provided not to surpass visual inspection for products and in workplaces. ANY’s Deep Learning Package solves these problems, and implements optimized image inspection into manufacturing workplaces in a true sense.
If your company has been disillusioned with image inspection up until now, deep learning provided by ANY is a tool that will bring about revolutionary changes for your company.
In conventional rule-based image inspection, for uses with variation in error decision and where identification with small differences in accuracy is unfeasible, enormous case studies must be input/created. This is a factor leading to long development periods and higher costs. There are other issues, such as changes in the specifications of the product or improvements to the production method necessitating reprogramming, causing the merits provided not to surpass visual inspection for products and in workplaces. ANY’s Deep Learning Package solves these problems, and implements optimized image inspection into manufacturing workplaces in a true sense.
Optimization for factory automation. We deal with the positioning, inspection, and classification of the components needed for decisions.
On-site training, and on-site assessment. Customers can make models by themselves.
ANY’s “Deep Learning” package is a completely stand-alone system. Our customers can conduct study and assessments themselves throughout easy-to-use interface, and can make models.
The created models can be output as a settings file, so settings can simply be copied over as a file when inspecting multiple products. Factories that produce small lots of many different products can benefit from the merits of image inspection using AI.
The created models can be output as a settings file, so settings can simply be copied over as a file when inspecting multiple products. Factories that produce small lots of many different products can benefit from the merits of image inspection using AI.
Customers can make their models in their companies. This can greatly reduce lead time and running costs.
Offering this as a package provides astounding cost performance.
Making it a package product specializing in machine vision allows us to offer amazing cost performance. This can not only greatly reduce workplace operating costs such as implementation costs, but also costs associated with model development and compatibility with new products.
ANY Deep Learning Package
| Main Functions | |
| Positioning | |
| Product Classification | |
| Defect Detection | 1. Teaching with only good products 2. Teaching with both good products and defective products (flawed points) |
| OCR (optical character recognition) | |
| Main Specifications | |
| Camera | Can be connected up to 4 to a GigE camera |
| Lighting | 1 light per camera; compatible with up to 4 in total |
| Input/output | Connection to PLC by SLMP connection (no need for a ladder for communication) |
| Other | Inspections combined with existing image processing are possible. The camera/lighting/PLC can be connected simply with an Ethernet cable. |
*Our Deep Learning Package uses “Vidi” from Cognex Corp.
Possible to combine 3D data and robotics
ANY’s Deep Learning Package is highly customizable and flexible. It is possible to conduct processing/analysis optionally using 3D data instead of images and taking images flexibly with robotics, we can implement an imaging system that is optimal for the objects to be inspected and the work environment. For more details, please contact us.