Research on Stages of AC corona discharge Based on Visible Digital Images Gray-level Co-occurrence Matrix

In terms of optical image digital processing of gas discharge, the morphological research is mainly engaged in the image of UV, which has relatively few studies on discharge state by using grayscale texture information of visible images. Thanks to the progress of photoelectric detecting technique, the visible image contains more abundant space structure information, which makes it possible to detect the characteristic of grayscale texture information. In this paper, the 50Hz AC high voltage power source was applied to the 15mm needle-plane gap under room temperature and atmospheric pressure. The discharge images in the time scale of seconds were obtained by a high resolution digital camera. We introduce the research method of gray-level co-occurrence matrix into the corona discharge image processing. Contrast, entropy, correlation and the expected value ratio of vertical (90 degrees) and horizontal (0 degrees) directions are used to describe the state of corona development. The calculation results show that the contrast and entropy can effectively reflect the corona pre-breakdown state, and the correlation can effectively reflect the initial state. The expected value ratio can effectively reflect the process of growth, suppression, and gap-penetration of the discharge in the positive half cycle, which has a good correlation with the oscilloscope electrical signals. The method allows us to take advantage of visible light images to study the stages of AC corona, which provides a new method of low temperature plasma diagnostics.