Breakthrough technology for complex inspection, component positioning, classification and OCR
Cognex's VisionPro ViDi suite is the first deep learning-based image analysis software designed specifically for factory automation applications. By combining artificial intelligence (AI) with VisionPro and Cognex Designer software, the Cognex VisionPro ViDi suite can solve complex applications that are too difficult, cumbersome or expensive for traditional machine vision systems.
This new method tolerates errors and unpredictable defects, can outperform even the best inspectors, and is most suitable for:
-
Defect detection
-
Material and material classification
-
Assembly verification and positioning of deformed components
-
Character reading, including distorted printing
Like human beings, powerful and fast
The Cognex VisionPro ViDi suite fully combines the professionalism and flexibility of manual visual inspection with the reliability, repeatability and performance excellence of the computer system, and everything is integrated into an easy-to-deploy window.
4 powerful tool sets
The deep learning algorithm of Cognex's ViDi suite is optimized for real-world industrial image analysis. Training and verification can be completed with a very small set of sample images and a short time. Choose between "Analysis" with a red background, "Classification" with a green background, "Locating" with a blue background and "Reading" tools with a blue background to solve applications that are highly supervised to completely unsupervised.
Find the feature
ViDi Blue-Locate (Blue-bottomed "Locate" tool) can find complex features and objects by learning with annotated images. Self-learning algorithms can find components, count transparent glass vials on conveyor trays, and perform quality control checks on kits and packaging.
Detect anomalies and lack of aesthetics
ViDi Red-Analyze (the "Analyze" tool with red background) can segment defects or other key areas by learning different changes in the appearance of the target area. ViDi Red-Analyze can identify scratches on complex surfaces, incomplete or improper assembly, and even weaving problems in textiles by learning the normal appearance of the object, including its significant but tolerable changes.
.png)
Categorize objects and screens
ViDi Green-Classify (the "classification" tool with green background) can distinguish different categories according to the marked image collection. Through training the acceptable error range, ViDi Green-Classify can identify products according to packaging, classify welds, and distinguish between acceptable and unacceptable anomalies.
Read text and characters
ViDi Blue-Read (the "reading" tool with blue background) can use optical character recognition (OCR) to decode severely deformed, skewed, and poorly etched codes. The pre-trained font library does not require programming or font training, can recognize most of the text, and implement it quickly and easily. You can also retrain this robust tool and adjust it to specific OCR application requirements.
Training and deployment are quite easy
Cognex's ViDi kit enables technicians to complete deep learning-based model training in minutes based on only a small sample image collection. After the application is completed, ViDi can quickly provide accurate results and store images for process control.
