Self-learning vision system supports automated measurements

New Products |
By eeNews Europe

The CV-X series now comes with automatic measurement setup. Users choose the type of measurement they want to perform and click the points they want to measure between. This automatically positions and sizes the inspection areas, optimises the settings to pick up on the correct features and creates the calculation. Targets are identified by Keyence´s new shape detection algorithm “ShapeTrax2” allowing for multiple stable inspections to be taken whatever the product position. Even with inspection accuracy down to sub pixel levels the processing speeds are far superior to the industry standard (even using 5 million pixels per image).
Once the camera is in position on the production line, running 30-50 good products will allow the camera to “learn” the acceptable manufacturing tolerances, checking for shape, colour and pattern. This is done by recognising the similarities between non-defective parts rather than differences in defective parts which greatly increases the stability of the inspection. 

The CV-X100 vision system is equipped with a suite of tools designed to make image capture more reliable. Among them is a tool to ensure that the image capturing conditions (camera position, lighting, focus) remain constant. The CV-X Series is also capable of using digital correction to remove the effects of curved or slanted surfaces due to camera installation and/or lens distortion. The result of this digital correction is an image identical to one taken from the perspective of a camera positioned right above the target. This function is very popular for robotic applications.
Traditional filters like binarisation or averaging as well as other much more advanced filters, such as HDR and shading correction can be used to enhance images from even the most complex targets.

Visit Keyence at


Linked Articles
eeNews Europe