Application of Computer Vision in warehouse and logistics

1. Inventory management:

Cameras are used to automate inventory management processes, such as barcode scanning, item recognition and item tracking. The benefits of real-time visibility of stock levels and improved accuracy in stock counting should be emphasized.

Modern logistics warehouses can no longer be operated efficiently without computer vision technology and automation solutions. Modern distribution centers with a maximum degree of automation are essential for maintaining global supply chains.

Where many work steps used to be carried out by a warehouse employee, automated processes with computer vision, robots, OCR/barcode readers and automated forklifts are now the standard. On the one hand, these greatly increase the cycle speed and, on the other, minimize errors caused by human error.

One of the most important components for these automation steps are industrial cameras, suitable lenses, the right machine vision lighting and a computer vision software for the respective application situation.

2. Quality control:

Cameras are used for quality control purposes, e.g. to check for product defects, detect damaged goods and ensure compliance with quality standards. Computer vision in warehouse and logistics can detect anomalies more efficiently than manual inspections.

3. Process optimization with computer vision:

Cameras are used to optimize warehouse processes, such as picking and packing. Computer vision software and image processing algorithms can analyze footage to identify bottlenecks, streamline workflows and improve overall efficiency.

4. Remote monitoring and management:

Remote monitoring and management using computer vision technology allows warehouse managers to access live camera feeds and analytics dashboards from any location to monitor operations, identify issues and make data-driven decisions in real time.

Industrial camera selection for warehouse logistics

In our customer application, a scanning tunnel is installed through which different parcels are transported directly to the warehouse via a conveyor belt. The parcels can be of different sizes.

The inspection task is to read barcodes on the parcels in order to automatically assign them to their corresponding storage location and adjust the stock in the warehouse software accordingly. An additional challenge with this inspection task is that the stickers with the barcodes do not run through the scanning tunnel at any defined point; they can be on the top or on one of the four sides of the parcel, the only exception being that the code cannot be on the bottom of the parcel. At the point where the image is captured, the package stops briefly.

First of all, we had to find a suitable industrial camera for this application. To be able to read a barcode reliably, we need to know how wide the thinnest barcode is. This line should be represented by at least 2 pixels to ensure that it is also reproducibly recognized by the image processing software.

In this application, a FOV ( field of view ) of 500x350mm would have to be covered by the camera, the thinnest line of the barcode has a thickness of 0.2mm. This results in a required resolution of 17.5MP or 5000x3500 pixels (0.2mm must be represented by 2 pixels-> 0.1mm/pixel = 5000 pixels on 500mm x 3000 pixels on 300mm).

A good choice for this task is 'MER2-2000-19U3M', a 20MP rolling shutter camera with a 1" Sony IMX183 sensor. Since the barcodes are black and white, a monochrome camera is sufficient for this application.

As the barcodes can be affixed to 5 different sides of the package (top, front, back, left or right), the customer decided to use one camera set for each viewing direction in order to reliably cover all possible viewing directions.

In principle, this work step could be simplified even further in the future by ensuring that the parcels are placed on the conveyor belt in a defined position on delivery. This would ensure that the barcode can only be on one side and a camera setup would only be necessary there in the scanning tunnel.

However, the customer was concerned that the entire process could stop regularly simply because the positioning was not carried out correctly when the parcels were delivered. This would then cause a complete stop of the conveyor line, and the costs of this interruption would quickly exceed the costs of the camera sets required to inspect all sides of the parcels.

Lens selection for IMX183

To determine the right industrial lens, we use our Lens calculator:

The MER2-2000 uses a 1” sensor with a resolution of 5496x3672 pixels at a pixel size of 2.4µm. Using a C-mount lens with a focal length of 16mm 'LCM-20MP-16MM-F2.8-1.1-ND1' at a working distance of 680mm results in a FOV of 547x366mm.

Machine vision lighting for warehouse logistics

In addition to the design of the camera and lenses, uniform and homogeneous illumination of the area from all sides is essential for this application. Our recommendation for this is machine vision lighting from our bar lights, which is available in various sizes and colors.

To protect the camera setup against external influences, the customer uses our metal protective housing (IP67). The particular advantage of these protective housings is their variability thanks to the modular housing design. Lenses of different sizes can be used. This gives the customer the option of using the protective housing even if the application changes, e.g. if the FOV or working distance changes.

Support for warehouse logistics with computer vision?

Computer vision technology offers a powerful solution for automating warehouse processes and optimizing logistics operations. By implementing camera systems with machine vision software, warehouses can achieve real-time inventory management, ensure quality control, streamline workflows, and enable remote monitoring.

As illustrated in the example of the scanning tunnel, careful selection of cameras, lenses, and lighting is crucial for successful implementation. With proper planning and the right components, computer vision can significantly improve efficiency and accuracy in warehouse logistics.

Would you like to implement a similar application? We would be happy to support you in selecting the right components for you warehouse and logistics automation. Please feel free to contact us!