DeepFlameS demonstrator with visualization of the segmented flames in the live image, and diagrams of the calculated characteristic parameters.
DeepFlameS project
Artificial intelligence for robust flame segmentation for the optimization of industrial combustion processes
Funding: Ministry of Economics, Labor and Tourism Baden-Württemberg
Duration: 2021 to 2024
Project consortium: Karlsruhe Institute of Technology (KIT), BASF SE, ci-tec GmbH
As part of the DeepFlameS joint project, artificial intelligence methods are used to optimize thermal process plants and thereby contribute to energy efficiency and the reduction of CO₂ emissions. The basis is provided by infrared camera images of industrial combustion processes in selected example plants. Deep-learning methods are developed and used to reliably segment flame bodies in image data and extract characteristic parameters for the current process state. In order to evaluate the suitability for daily use, the developed methods will also be implemented in an industrial-suited software demonstrator.
The DeepFlameS joint project is funded within the framework of the InvestBW program.