This project focuses on researching object recognition in image files and developing a concept for detecting objects in existing or self-generated datasets. The goal is to digitize and optimize the logistics industry through the implementation of AI solutions. Specifically, automatic measurement of cargo items in logistics centers is aimed at, to make truck loading more efficient and save space on the loading area.

Currently, loading planning is often based on the assumption that cargo items are simple geometric shapes such as cubes or cuboids. Deviations from these shapes, such as protruding parts, are not taken into account, which can lead to inefficient use of cargo space. However, by applying artificial intelligence, cargo items can be measured directly during the unloading process, and irregular dimensions can be accurately captured.

Since specific logistics operations are strictly confidential, obtaining video or image material showing the unloading process of a truck poses a major challenge. Therefore, a simulation environment has been developed to simulate various unloading scenarios, allowing the training and optimization of object recognition algorithms.

• You can find more information on the Showcase website.

Students:

• Raphael Link

• Sascha Binkert