During Paschen tests of fusion reactor magnets, it is essential to establish an insulation monitoring system to locate insulation faults. To provide reference data for the design of such a system, this paper first establishes a Paschen discharge experimental platform based on visible light imaging and then develops a discharge diagnostic system. Given the large accumulation of monitoring video files during discharge experiments, a discharge frame detection algorithm (DFD), based on histogram analysis and discrete point detection, is designed to improve the detection speed and accuracy the of discharge frames.

After testing the Paschen discharge experimental platform, it was found that the gas breakdown characteristics in nonuniform fields still follow the basic principles of Paschen’s law. Moreover, the change in plasma luminosity with vacuum pressure correlates to the behavior described by Paschen’s law. Finally, performance testing of the DFD algorithm shows that DFD achieves a F1 score of 99.4%, surpassing traditional detection algorithms based on edge detection and Fourier transform. The detection efficiency of DFD is 13 times and 21 times higher than these traditional methods, respectively.

The results of the Paschen discharge experimental platform tests demonstrate that monitoring gas discharge phenomena during Paschen tests with a visible light camera is highly feasible. The discharge diagnostic system, with its efficient and precise performance, can quickly capture nearly all discharge images, providing valuable experience for the technical development of magnet insulation monitoring systems.