Deprecated: stripos(): Passing null to parameter #1 ($haystack) of type string is deprecated in /home2/ldhvisio/public_html/wp-content/plugins/stackable-ultimate-gutenberg-blocks/src/init.php on line 191
Deprecated: stripos(): Passing null to parameter #1 ($haystack) of type string is deprecated in /home2/ldhvisio/public_html/wp-content/plugins/stackable-ultimate-gutenberg-blocks/src/init.php on line 202
Deprecated: strpos(): Passing null to parameter #1 ($haystack) of type string is deprecated in /home2/ldhvisio/public_html/wp-content/plugins/stackable-ultimate-gutenberg-blocks/src/fonts.php on line 42
Deprecated: strpos(): Passing null to parameter #1 ($haystack) of type string is deprecated in /home2/ldhvisio/public_html/wp-content/plugins/stackable-ultimate-gutenberg-blocks/src/fonts.php on line 42
Deprecated: strpos(): Passing null to parameter #1 ($haystack) of type string is deprecated in /home2/ldhvisio/public_html/wp-content/plugins/stackable-ultimate-gutenberg-blocks/src/jetpack.php on line 56
Deprecated: Creation of dynamic property OMAPI_Elementor_Widget::$base is deprecated in /home2/ldhvisio/public_html/wp-content/plugins/optinmonster/OMAPI/Elementor/Widget.php on line 41
LDH-Vision turn-key inspection systems are designed to meet the specific needs of each customer. Example parts for which an LDH solution is particularly well-suited are those with cylindrical geometries produced by the powered metal industry. For these, requirements frequently include:
- Production rates between 500 and 30,000 parts per line per shift.
- Full inspections of all four surfaces (top, outside, inside, bottom) of rings, gears and bushings.
- Inspection times on the order of 1 second per part (3,600 parts per hour).
- Qualitative classification of defects including heavy coatings, damaged gear teeth and scratched surfaces.
- Storage and display of results to guide both short-term and long-term production line process control.
LDH-Vision systems are powered by deep learning – a form of Artificial Intelligence. Deep learning classifiers:
- Use deep convolutional neural networks (CNNs) which allow for extremely complex mathematical models.
- Learn by example to distniguish between good and bad parts from images of each.
- Improve classification accuracy by iteratively correcting for identification misses.
- Can identify subtle defects which are not recognizable by conventional machine vision.
- Capture the ability of expert human inspectors to recognize defects and shortcomings.