Abstract: In this paper, a high-speed camera is used to record the gas discharge information and obtain high-resolution spatial and temporal information. It is found that the discharge pattern is a tree fractal structure, and the fractal dimensions of a single discharge point and multiple discharge points are 1.64 and 1.71, respectively. Through the intelligent contour recognition and centroid marking of a large amount of image information obtained by self-written programs in Python, the location distribution information of discharge points of 5911 images with a time interval of 0.5 ms are obtained. The results show that the location probability of discharge points in space is close to Gaussian distribution, and the time interval distribution of its occurrence is close to exponential distribution.

Key words: random phenomenon, gas discharge, image processing, intelligent identification