AI in Manufacturing Industry – Leanest Scanner 2.0

Technology

In our quality detection software solutions for manufacturing industry, we have focused on AI solutions based on image recognition. We have specialized in image processing algorithms (image recognition) and machine learning.

Technologically we are using cameras and 3D lasers for capturing the material, applying additionally detection software that has been developed to artificial intelligence.

All solutions are built by us according to each customer unique needs and the choice of technology is made according to the conditions and requirements.

Leanest Scanner 2.0

We have developed Leanest Scanner as a standard solution, significantly shortening project implementation time.

Leanest Quality Scanner employs artificial intelligence during operation, specifically trained on customer product and material-based data to detect visually identifiable quality deviations in products.

The device is designed to assess quality as the product moves longitudinally and is positioned between two conveyors. The compact device has a footprint of 550x1250mm.

Operator Workstation

The monitor can be the customer’s preferred display. As a standard option, we use a 15.6″ touchscreen monitor. The orientation of the monitor can be adjusted 180 degrees based on the position from which the scanner will operate and where the material’s zero line is located. For distance and angle adjustment, we use a standard VESA mount.

Material

Depending on the cross-section of the material and the customer’s preference, we can design the placement and number of cameras for product inspection. We use a motorized stage to ensure the correct focal distance from the material at all times.

For inspecting the cross-section of the material, we can add appropriate lasers. We use red-colored lasers (also possible in blue) capable of operating at up to 10,000Hz.

Inspection can be carried out from all four sides, searching for characteristic objects on each side separately.

The aperture through which the material must pass is standardized at 260 x 100 mm, but depending on the product, the device can be adjusted.

Technical Specifications

Cameras: Japanese industrial cameras with industrial-grade lenses capable of operating up to 26kHz.

Lighting: Lighting designed for industrial machine vision systems from a European manufacturer.

Laser: Lasers from a German manufacturer operating at frequencies up to 10kHz and providing measurement accuracy up to 0.01 mm (depending on material and production line).

Material dimensions: 260 x 100 mm.

For transmitting quality decisions to the production line, digital signal, Modbus, or other Ethernet protocols can be used.

European and US suppliers have been used to ensure the device build quality. This is important for both – overall assembly and for technical support quality. The device is assembled and manufactured in European Union, Estonia.

Potential Use Cases

We apply artificial intelligence technology in areas where conventional ways could not keep up with the rising quality standards by the customers, e.g., in quality control jobs in the factories.

Some examples of potential use cases:

Automatic sorting of material in the wood industry. Read more…
Quality control in the metal industry
Plastic sorting in the waste industry. Read more…
Input material control in the food industry
Quality control of products in the electronics industry.
Quality control of packages/labels.

Based on the needs of each customer, we choose the best suitable AI solution.

Benefits of AI

Solutions based on artificial intelligence ensure stable and effective quality control in production. We eliminate errors and quality instability caused by the human factor.

AI works fast. Depending on the chosen solution, it is possible to issue a decision in 1-100ms. With such speeds, it is possible to increase line speeds significantly and achieve greater productivity.

We replace routine manual work with an automated process and significantly save resources. Until now, this role has been fulfilled by human labor, but increasingly expensive labor makes production gradually more expensive.

Trained AI algorithm can make better-than-human product decisions. The algorithm does not get tired and always makes decisions on a solid basis.

During the training suitable examples need to be provided and the learning process and finding connections is done by the algorithm itself. As new samples are added, there will be constant replenishment needed.

With the help of AI and corresponding devices, work processes can be executed safely that could otherwise potentially endanger the health of employees (e.g. sorting waste processes).

In order to stay competitive in a globalizing world, modern production needs a functioning quality control system. Today's quality control is essentially a software solution. With our experienced team, we bring these technologies to the industry and provide complete solutions for quality control.