The erosion of the switching contacts caused by vacuum arcs during breaking operations determines the durability of vacuum circuit breakers. The scale of this erosion is depended on strength and movement of the vacuum arcs. These arcs consisting of metal vapor are an intense radiation source. The radiation distribution of the arc provides information about arc strength and movement. This distribution could be investigated optically in a demountable vacuum chamber. Based on this radiation information about the position and mode of vacuum arcs are identified. Typical modes are diffuse, constricted or jet modes. A high-speed camera records the radiation intensity of high-current switching tests with transversal magnetic field (TMF) contacts from two rectangular perspectives. The recorded images include the arc information about the mode, the position and the velocity. To achieve reliable results independently from the individual observer within a reasonable time out of a high number of switching tests, a software-based solution is required. For this purpose the development for an automated evaluation routine using LabVIEW is presented. In a first step, the recorded radiation intensity of the images from an average switching test with a 10 ms sinusoidal half-wave have been digitized. All measured data like video images, current flow and arc voltage, contact stroke are digitally synchronized to a common timeline. In further steps, the images are digitally processed to prepare an automated arc mode recognition and extract specific features. The developed software solution is tested with images of high-current switching tests up to 31.5 kA (RMS) using TMF contacts with a diameter of 68 Millimeters. The experimental setup for the arc localization and the gradual processing of exemplary images is presented in this paper.