Recognition systems and image exploitation

Recognition systems are systems and procedures that extract higher-level information and knowledge about entities in the real world from incoming data or assist humans in doing so.

The main focus of IOSB’s activities lies on researching and developing recognition systems based on imaging sensors. We employ imaging sensors covering the visual as well as the IR spectral ranges, which are processed as single images or video image sequences. The real-time aspect and process optimization for lightweight hardware and application-specific solutions play a major role in our work. Our recognition methods are primarily deep learning  approaches, which are adapted or developed specifically to fit the task. We train and evaluate our models on our own, extensive deep learning server infrastructure. While we have a wide range of annotated datasets already available for training, we have also established an efficient process for curating and annotating additional training data using several tools and workflows developed specifically for this process.

Fields of application

In Airborne Image Exploitation, Fraunhofer IOSB is able to detect and track persons and vehicles automatically and in detail with deep learning methods in Wide Area Aerial Surveillance. Optronic sensors are also used in Airborne and Ground-Based Image Exploitation, while sea-based sensors and CNN-based methods are used for ship detection and recognition.

We offer privacy-compliant, intelligent systems for the detection, recognition and activity recognition of persons. This is used, e.g., for person density estimation at mass events, for person recognition based on descriptions or in AI-supported intelligent video surveillance, which reports only dangerous situations with trained activity recognition.

The detection and defense of illegal drones is a major challenge, especially for airports. After initial detection by radar or radio, Fraunhofer IOSB additionally uses video sensor technology. Drones are accurately distinguished from decoy objects such as birds. With our detection and tracking methods for drone tracking, drone types can also be recognized.

For the inspection of bulding infrastructures, we are developing deep learning methods that detect changes to objects using airborne change detection with video sensor technology. The information is condensed so that only relevant changes such as corrosion effects or missing parts are displayed. 

Education and training of the image exploitation specialists serve to ensure the competence to act and contribute to an effective and efficient use of detection and image exploitation systems. For this purpose we develop intelligent, adaptive learning systems including interactive animations, serious games and gamification methods.

With an AI-based combination of two- and three-dimensional image information, we provide an accurate capture of the environment for the navigation of mobile systems. With our mapping and obstacle detection, robotic systems, for example, can move accident-free in their workspace, even off-road on unstructured terrain.