Intelligent imaging

The majority of our work is focused around 4 areas, which is divided under two clusters Imaging4Autonomous Systems and Imaging4Decision Support. The cluster Imaging4Autonomous Systems works is concerned with Robotics and Inspection. Imaging4Decision Support works in the field of Defence and Security. Besides these 4 areas, we also do projects in the space, health and industry domains.

Imaging4Autonomous systems

The Imaging4Autonomous Systems clusters specializes in building vision systems for the exact needs of our customers. Whether their needs are to be able to ‘see’ at 120 km/h on the highway, or 12 meters under water, we will help them construct the right combination of sensors (sensor suite design). Testing these systems is the fun part; this is where we get to take our prototype system, put it in or on a vehicle and start gathering data. Often, these raw data are not our end goal. Through image processing and computer vision techniques we convert raw data to actionable intelligence.

The next step within our expertise is to put the sensors on robotic platforms and use the sensor information to make them operate remotely or more automatically. For this, we try to extract only relevant information from the overload of image data from the sensors, in which we are experts. Another expertise of our group is to use the sensors to reconstruct the 3D world. With the goal that robots and humans can understand, reason about it and act in it. Also, nowadays our 3D reconstruction and sensor suites are often streamed and visualized in Virtual Reality (VR) for various purposes.

Imaging4Decision Support

The Imaging4Decision Support specializes in image and video analysis for supporting humans in their decision making within Security domain, Defence domain and other domains.

Firstly, this clusters develops documentation authentication technology for border guards and immigration services, to support experts in the assessment of official documents. By developing smart image processing techniques, we can ease this challenging and time consuming task and not only make their work easier, but also more efficient. Within this cluster we also focus on processing for camera footage in surveillance and security applications. Security operators are responsible for observing large amounts of camera streams, and to help them detect threats early and on time we develop behavior analysis, anomaly detection and person re-identification technologies. Finally, within all these developments we also put effort in anonymization and privacy-enhancing techniques, so that these sensitive data and models can be shared.

In the Defence domain, the cluster focuses on supporting almost all disciplines within the armed forces with extracting relevant information from images. Using expert knowledge in the field of AI, state of the art technologies can be applied in diverse projects within the Defence domain.

For the Royal Navy we recently developed an application which is able to automatically identify military ships appearing on the horizon. In an airborne context we are working on automatic processing of WAMI images, which enables automatic detection and tracking of then thousands of vehicles at the same time. These solutions increase situational awareness while reducing the cognitive load of operators and analysts. We are also involved in several projects about the interpretation of image sensors in autonomous systems. We also have projects on the topic of image enhancement technologies, such as the development of new turbulence correction technologies. In these projects we support the armed forces with both hands on proof of concepts as well as conceptual research and evaluations.

Collaboration outside Intelligent Imaging

Our department is a well-known player in the domain of image processing and collaborates on this topic with a diverse set of partners. Within TNO, we work closely together with other expertise such as Data Science, Human-Machine Teaming and Intelligent Autonomous Systems. Outside of TNO, we have an even larger set of partners. The Dutch Ministry of Defense is one of the biggest customers of Intelligent Imaging.

Within the Security Domain we work together with almost all national organizations as well as some international institutions. These include the National Police, the Royal Marechaussee (KMAR), Immigration and Naturalization Service (IND), Custodial Institutions Agency (DJI), FRONTEX and European border guards, Europol and European law enforcement agencies and other such as NATO, EDA and DARPA. Finally, we also work with universities (University of Twente, TU Delft, University of Reading), public institutions (Department of Waterways and Public Works and NLR) and commercial organizations, such as manufacturers of optical equipment and robotic industrial partners.