Artificial intelligence of defect detection of polymer products in production line

Abstract

The invention discloses a method for detecting defects in at least one article, the method comprising the steps of: acquiring, via an image acquisition module (104), at least one image of an article in a production line; processing the acquired image via artificial intelligence (106) to determine whether any defect is present in the item; discharging defective items from a former in the productionline via a purge system (105); and updating the result to the artificial intelligence (106) for subsequent detection. The artificial intelligence (106) trains its dataset using a neural network by performing an iterative training process to create at least one data classification from a plurality of pre-stored images; and generating a weight formula according to the data classification so as to evaluate the data of the acquired image.

Description

technical field

[0001] The present invention relates to a detection system for defects and a method thereof. More specifically, the present invention relates to an artificial intelligence system and method thereof for detecting defective gloves in a production line. Background technique

[0002] In modern industry, latex gloves are usually produced on high-speed production lines to meet the demands of supply chain management. Manual defect detection of products often results in inaccuracies due to human biological constraints of performing repetitive tasks over time. In order to solve this problem, various automation systems have been developed by implementing sensors such as vision, laser, or ultrasonic waves to detect defects of products to improve productivity.

[0003] For many years, there have been few patented technologies related to automatic defect detection systems. An example of a machine vision-based glove tearing and oil stain detection method is disclosed in...

Images