In order to provide redundant data support for attitude angle calculation of UAV, a skyline detection algorithm is introduced. A skyline extraction area and a skyline detection gradient threshold are determined, the skylines of each column in the image are detected, the detected skyline singular point values are eliminated, and the included angle between the skyline and the horizontal line of the unmanned aerial vehicle image is calculated as a data source for attitude calculation. The results show that the intelligent image processing module designed by this method can effectively detect the skyline of the image in real time, calculate the attitude angle of the UAV, and control the UAV to fly stably.