Self-calibration method and device for binocular camera of unmanned aerial vehicle

The invention relates to a self-calibration method and device for a binocular camera of an unmanned aerial vehicle, relates to the technical field of computer vision, and solves the problems that the calibration process of an existing binocular camera of the unmanned aerial vehicle is relatively tedious and the user experience is influenced. The self-calibration method for the binocular camera of the unmanned aerial vehicle comprises the following steps: step 101, acquiring a binocular image; 102, extracting a first angular point from one of the binocular images by using an accelerated segmented test feature point extraction algorithm; 103, extracting a second angular point from the other image in the binocular images by using an optical flow method, wherein the second angular point and the first angular point form a binocular matching point; and 104, constructing an epipolar error objective function according to the first angular point and the second angular point, optimizing the epipolar error objective fun