Revolutionizing Quality Control in Canning Lines: Advanced Can Inspection with Deep Learning

In the canning and beverage industry, where speed and safety are critical, automated visual inspection has become an indispensable tool to guarantee the quality of the final product. In this context, we present a state-of-the-art solution based on artificial vision systems with deep learning technology, capable of inspecting both empty cans (Empty Can Inspector – ECI) and full cans (Contamination DL) with maximum precision and at industrial speeds.

VisioInspect 2.0 + AI Technology: Intelligence that Learns to See

Both the ECI (Empty Can Inspector) and Contamination DL systems use the powerful VisioInspect 2.0 software, developed by E2M COUTH to process high-resolution images using artificial intelligence algorithms trained on thousands of real examples. This combination allows overcoming the limitations of classical methods, achieving a more robust and flexible detection without the need for rigid parameterization.

ECI – Intelligent Inspection of Empty Cans

The ECI system developed by E2M COUTH operates at speeds of up to 120,000 units/hour and is designed to inspect empty cans of different formats (height from 35 mm to 120 mm and diameters from 60 to 100 mm). Its capabilities include:

∙Body ovality: deformation detection with an accuracy better than ±0.5 mm.

∙Analysis of the “flange”: Verification of width, presence of peaks, bends and dirt or opacities from 1 mm².

∙Internal contamination: Identification of foreign objects from 1×1 mm on the internal surface.

∙Absence of varnish: Recognition of uncoated areas inside the can.

∙Structural shocks: evaluation of impacts that affect the shape or integrity of the package.

Thanks to its optimized optical and lighting architecture, the system achieves glare-free images, even on mirror-finished cans.

Contamination DL – Full Can Inspection with AI Accuracy

Designed for production lines up to 70,000 units/hour, the Contamination DL system developed by E2M COUTH is focused on guaranteeing the quality of the already packaged product. Its functions include:

∙Ovality control: automatic rejection of cans with diameter variations greater than 0.5 mm.

∙Detection of visible contaminants on the product: Analysis of impurities, dirt or foreign bodies larger than 2×2 mm.

∙Impact assessment: Identification of visible damage or damage affecting the geometry of the container.

With a single high-resolution camera and a custom-developed illumination system, the Contamination DL obtains accurate images of the content without optical interference, even at high speed.

Why bet on Deep Learning?

The use of AI not only improves the detection of unstructured defects, but drastically reduces false rejections, adapts to new defects with ease, and enables system retraining for product or process variations. Unlike traditional approaches, where each new case requires adjustments, the deep learning models developed by E2M COUTH learn by direct visual experience, just as an expert human inspector would.

Advantages

∙Reliable inspection at industrial speeds (>100,000 u/h).

∙Adaptability to multiple formats and materials.

∙Reduction of wastage and costs due to false detections.

∙Improved traceability through automated analysis of each unit.

∙Easy integration with existing production lines.

Take your quality control to the next level!

Incorporating machine vision with deep learning into your production line is not just a technological improvement: it is a commitment to operational excellence, consumer safety and global competitiveness. Contact our team for a personalized demo and discover how to transform your inspection process.