By Chris Gilmet, Equipment Engineering Manager at Acquire


Traditional machine vision systems fall into the category of Machine Learning, where engineers and technicians teach vision systems what to look for like matching barcodes, finding patterns, and measuring edges. Our team is taking things a step further with Artificial Intelligence and Deep Learning, a new concept in the realm of machine vision where systems are taught to perceive images much like humans do by recognizing and adapting to varying features. Deep learning is a branch of machine learning based on a set of algorithms that model high level abstractions in data and has been characterized as a rebranding of neural networks.

Deep learning opens the door to numerous inspection applications and solutions that were previously unsolvable within high speed, environmentally challenging industrial environments.  Here are three huge advantages to utilizing deep learning for your machine vision application:

1. Deep Learning can eliminate unnecessary costs.

Recalls are expensive. In the Food and Beverage industry alone, a recall costs a company an average of $10 million in direct costs. Deep Learning can detect more subjective defects that are difficult to train such as minor product labeling errors like incorrect fluid ounces or region that would relate to a significant recall.

2. Deep Learning identifies defects that would otherwise be difficult to detect.

When consistent images are challenging due to ambient conditions, product reflection, or lens distortion, Deep Learning can account for these types of variations and learn interesting features to make your inspection robust.


3. Product variation can render traditional Machine Learning impractical.

Machine Learning can’t consistently inspect irregular shapes and patterns that do not have repeatable edges like the example below. Deep Learning successfully inspects these items.



To learn more, reserve your spot for my webex on March 8.

Pin It on Pinterest