COGNEX VIDI – DEEP LEARNINGIMAGE ANALYSIS SOFTWARE
Cognex VisionPro ViDi is image analysis software based on so-called "deep learning“. The SW platform comes under this term and is based on the principle of neural networks. In practice, the term artificial intelligence is often used. The ViDi tools described below make it possible for us to solve a new area of problems that previously could not be solved.
COMBINATION OF ARTIFICIAL INTELLIGENCE WITH INSTRUMENTS
Due to the combination of artificial intelligence with VisionPro tools and Cognex Designer software, VisionPro ViDi can solve problems that are too complex for classic machine vision.
This system is ideal for detecting defects, classifying textures and materials, verifying assemblies and finding deformed parts, as well as reading difficult-to-read text.
FOUR VIDI TOOLS SOLVE THESE PROBLEMS,THEY CAN BE USED SEPARATELY OR IN COMBINATION.
BLUE – LOCATE
The blue tool finds complex characters and objects from learned images. The tool can be used to find and count given objects, to detect individual components in assemblies, to check the complete package, and as a fixation for other tools.
RED – ANALYZE
The red tool is used to find defects, anomalies and aesthetic defects. This tool can detect defects on surfaces with different patterns and textures; normally a problem for classic machines vision systems.. Suitable applications include detection of scratches and defects on complex surfaces, inspection of defects on patterned fabrics, inspection of correctness and integrity of complex prints, detection of damaged or incomplete products or components. This tool can be used to measure an area by segmenting the selected character, and it can use dynamic masking to obtain a region for the use of other tools.
This tool can be used in two modes - unsupervised and supervised. The difference between these variants is that with the unsupervised option only a set of OK and NOK images is selected and learning with the ViDi Red unsupervised tool is performed. The supervised option has another step, which is when the programmer intervenes in the process and marks the required errors and defects and then learns with the ViDi Red supervised tool.
GREEN – CLASSIFY
The green tool is used to classify objects or entire images. This tool can be used to distinguish several types of products, identify products based on their packaging, check the integrity of packaging with many types of components and classify the defects found into different categories.
BLUE – READ (OCR)
The blue reading tool is suitable for reading characters and texts. The tool can also read deformed, scratched or poorly printed characters or characters on an inhomogeneou surface that are difficult to read. The tool also includes a library of pre-learned characters (OCR) through which the tool identifies most texts without the need for further programming or learning additional letters. The system can also be re-learned according to the specific requirements of the OCR application.
