Conveyor belt inspection
Identify and quickly repair flaws in working conveyor belts
Prodotti correlati
Ambiente di programmazione grafica per l'analisi di immagini industriali basata sul deep learning
Software di visione basato sul deep learning con tutta la potenza di In-Sight ViDi
Conveyor belts are subject to significant and constant wear, whether distributing packages in a fulfillment center, filling snack food packages, or packaging chicken parts. Many are metal with a variety of weaves, links, and rods specific to uses such as freezing, baking, washing, sorting, and cooling. Heavy use and resulting wear results in poor tracking, slack tension, or missing or bent links.
Unaddressed maintenance issues can damage products, resulting in waste and potential safety hazards, or even require an unexpected shutdown of the line, with resulting production delays. Preventive maintenance helps keep the line operating effectively with minimal downtime.
Keeping the belt under continuous observation during operation allows for the detection of current or potential defects. When the line shuts down daily, the defective parts can be located and replaced.
The variety of bends, rips, and other damage to the wire or metal links, rods, and other conveyor parts is so wide that it is impossible to program conventional machine vision to identify them all. Conveyor belts are often filled with debris, bits and pieces of product, blown dust, and other visual confounders, making conventional machine vision unreliable even for the defects they are programmed to detect.
Cognex Deep Learning makes continuous, accurate inspection of working conveyor belts routine. The defect detection tool trains on a set of images of functional conveyor belts, within acceptable levels of wear, and then flags any wear or damage outside of the acceptable parameters, regardless of confusing debris.
After defect detection, the classification tool enables process improvement. By training on image sets of both acceptable and damaged conveyor belt parts, it learns to classify specific types of damage or wear, providing users with the information they need to identify the source of excessive wear or damage and make modifications to reduce or eliminate it.
