Deep learning with Cognex ViDi Suite
Artificial intelligence for industrial image processing
Whenever complicated, time-consuming inspection procedures for manufacturing processes are required and where programming with traditional image processing would be very cost-intensive, the deep learning technology can be a useful add-on. Cognex ViDi is the first deep learning-based software, which was especially developed for such complex fields of application in factory automation.
Cognex ViDi offers the first ready-to-use deep learning software for industrial image analysis. The Cognex ViDi Suite is a tried and tested, optimized and reliable software solution, which is based on a cutting-edge set of algorithms in machine learning.
Traditional image processing stems from Machine Vision, hence the automation and inspection in standardized batch and series production. The advantage is a set environment with very precise knowledge about the kind of products or components that are being inspected. It is therefore possible to define a clear rulebook for an explicit Yes/No classification.
However, fixed rule concepts are no longer applicable for the factory of the future: Industry 4.0 is based on flexible, self-adapting production systems. Machine learning concepts are the next evolutionary step in image processing and do therefore require less rigid rulebooks, but are rather based on pre-processed data. Patterns or CAD models form the basis for the classification and the machine completes a training phase to acquire the knowledge.
In this training process it absorbs large amounts of data and learns how to separate "good from bad" by applying a manual assessment. The mathematical concept of a multidimensional feature space is used to define areas without reciprocal overlap, which enable a better classification of objects.
Solve sophisticated applications
Cognex ViDi Suite deals with solving difficult inspection and classification problems. The result is a high-performance and flexible program for numerous sophisticated fields of application in image processing. Successful use cases are to be found in the pharmaceutical, medical technology, automotive, textile, print, logistics and watch industry.
A deep learning-based software adapted to automatic manufacturing can:
- Solve image processing applications, which would be too difficult to program with rule-based algorithms
- Compensate distracting backgrounds and bad image quality
- Maintain and retrain applications at plant level.
- Adapt to new examples without the need to reprogram core algorithms
- Be operated by employees who are not image processing experts
- It is better than the best quality inspectors.
- Simple usage.
- Is able to manage even the most difficult tasks.
The integration of the Cognex ViDi deep learning image analysis in your company has a lot of advantages:
- Mechanical engineers expand their inspection solutions, reduce the development time drastically and increase their time-to-market speed.
- You benefit from a cost-efficient and reliable automatic quality control, which is better than any human quality inspector.