Electric arc detection method, device and system based on motion characteristics and storage medium

The invention relates to the field of safety detection of photovoltaic power stations, and discloses an arc detection method, device and system based on motion characteristics and a storage medium, which are used for improving the accuracy of fault arc detection. The arc detection method based on the motion characteristics comprises the following steps: acquiring a real-time monitoring image of a target area, determining the real-time monitoring image as an initial image when pixel points with pixel values higher than a threshold value appear in the real-time monitoring image, acquiring a preset number of monitoring images at a preset frequency from the initial image to obtain a target image sequence, determining a target feature attribute and a target analysis function matched with the target image sequence based on a preset parameter extraction strategy library, processing the target image sequence based on the target feature attribute to generate a target motion oscillogram, calling the target analysis fun