In order to further improve the productivity and quality of laser-based welding processes, process regulation is in many cases unavoidable. Because of the complex and highly dynamic processes, such regulation must be based on ultra-rapid optical image processing. CNN camera systems achieve major data reduction through intelligent neuron networking and therefore make image processing in the multi-kilohertz range possible. A technology enabling rapid, camera-based process regulation to be carried out has emerged from the joint project ACES between the Forschungsgesellschaft für Strahlwerkzeuge mbH (FGSW), the charitable An-Gesellschaft of the Institute for Radiation Tools (IFSW) at the University of Stuttgart and the Fraunhofer Institute for Physical Measurement Engineering (IPM) in Freiburg. Using this innovative technolgy it is possible to regulate laser welding processes regardless of material thickness, speed and other values influencing the process via automatic adaptation of the laser output with a regulation rate of up to 13 kHz on penetration welding.