Unmanned aerial vehicle visual landing runway start line detection method based on angular point features

The invention discloses an unmanned aerial vehicle visual landing runway start line detection method based on angular point features. The method comprises the following steps: carrying out graying processing on an acquired runway image; smoothing and blurring the runway image by using a Gaussian filtering method; carrying out edge detection on the runway image by adopting a Sobel operator; performing expansion and corrosion morphological processing on the detected edge features; establishing a dynamic mask according to the runway start line detection result of the previous runway image; carrying out Harris corner detection on the runway image; and performing spatial filtering on the detected angular points to obtain a runway starting line detection result. According to the method, the detection accuracy and retrieval efficiency of the initial line of the autonomous landing runway of the unmanned aerial vehicle can be improved, so that autonomous accurate navigation of autonomous landing of the unmanned aerial