Fraunhofer Institute for Digital Media Technology IDMTArtificial intelligence for improved acoustic quality control
Due to the increasing degree of automation and the growing complexity of manufacturing processes and products, smart methods are in demand in quality assurance, maintenance and repair. With AI-based acoustic monitoring and acoustic event detection, the business unit Industrial Media Applications offers a reliable and cost-effective method that can be used along the entire value chain for quality assessment. AI-based acoustic monitoring is also suitable for traffic or construction site monitoring, protection of critical infrastructure or biodiversity measurements. Depending on the application, there are specific requirements, on the one hand regarding energy efficiency of the detection technology, on the other hand regarding security and data protection. The business unit develops individualized solutions tailored to the customer's problems.
The research areas include improving the robustness of the analysis algorithms, the fusion of multimodal sensor data and the compression of trained AI models in order to make them available for energy-efficient analysis on mobile devices with limited computing power. Acoustic quality assurance with AI - simple, energy efficient, trustworthy.