Certified measurement with computer vision

Another focus of the project is to be able to use camera systems in combination with computer vision methods for high-precision measurement tasks in different application areas, with certification and calibration guaranteeing requirements and accuracy limits for the corresponding applications. In this context, a concept for the certification of camera-based measurement systems is to be developed. In the course of digitalization, cameras are now used in a wide range of devices for inspection and environmental detection. Examples include borescopy for the maintenance of hard-to-reach equipment such as turbines, endoscopy/microscopy in medicine, environment detection by drones or autonomous vehicles, and for quality assurance and detection of safety areas in manufacturing. In this context, digital acquisition together with suitable image analysis increasingly enables measurement of the environment beyond pure visual inspection - for example, as a haptic-auditory-visual measurement system (using sensor fusion). In this way, damage can be quantified during maintenance tasks, implants and prostheses can be manufactured in a patient-specific manner, or the quality of manufactured components can be checked. In order to ensure a broad application of the methods and approaches to be developed, the field of camera measurement technology must thus also be explicitly considered. However, especially for a broad use of cameras for measurement tasks, reliable sensor systems that meet the requirements for measurement accuracy and interchangeability are still missing. The project therefore aims at the conception of a certified camera system with guaranteed properties.

The HHI can draw on diverse preliminary work on wireless communication for data exchange in industrial applications, including ICT architectures for secure, latency-optimized communication and data processing in massive sensor networks (IC4F project), highly reliable real-time communication for industry and commerce (SEKOM project), for discrete manufacturing (TAKTILUS project) and for (ad hoc) networking of agricultural and construction machinery (AMMCOA project), highly synchronous real-time transmission (LIPS project) and secure radio-based measurement and control technology in district heating networks (OrtoFern3D project). In addition, it has a long-standing expertise in camera-based analysis and measurement of scenes and objects. In the M3D project, computer vision methods for 3D modeling of spare parts and in the 3DInMed project, methods for depth-based measurement of objects using 3D endoscopy/microscopy were developed. Further topics are: Image-based measurement methods for robotic channel inspection (AuZuKa), image-based methods for the measurement and positioning of objects in industrial manufacturing (EASY COHMO), image-based 3D cavity reconstruction with stereo endoscopes for clinical and industrial applications (3DHoropter) and high-gen