Abstract: In order to solve the problems of slow inspection efficiency and poor inspection accuracy of traditional power systems,an intelligent inspection technology for UAV autonomous driving based on cloud control coordination is proposed. A cloud control collaboration platform is built. The cloud server is connected to the intelligent inspection system through the intranet,and the UAV inspection multi-thread positioning module is designed,the coordinate information of the image coordinate module is received through the PC,and the display module realizes data sharing through the cloud platform. The coordinates of the center point of the image are divided and marked,and the Gaussian kernel function is used to improve the concentration of pixel distribution. Finally,the hovering point of the drone is drawn up,and the inspection route image obtained through the cloud control collaborative platform is detected to judge whether there is an abnormality,so as to realize the intelligent inspection operation of autonomous driving.Experiments have proved that the proposed method can accurately identify the edge of an image with an initial image pixel of 574*574,the maximum horizontal error is 0.5m,and the maximum vertical error is 1m. There is no deviation or jitter in the movement of the inspection route,which can provide certain technical support for the reliable and safe operation of the power grid.