Unmanned aerial vehicle field GNSS-free positioning method based on image matching

The invention provides an unmanned aerial vehicle field GNSS-free positioning method based on image matching. The method comprises the following steps: performing orthographic processing on an unmanned aerial vehicle image by using a vision-inertial odometer method in an image orthographic module; the coarse positioning module performs preliminary positioning on the unmanned aerial vehicle image based on image matching; and the fine positioning module is used for accurately positioning the unmanned aerial vehicle image by using Monte-Carlo positioning. The main objective of the invention is to develop a positioning algorithm which does not depend on GNSS for long-distance flight but only depend on a monocular wide-angle camera, the positioning method is practical in the case where GNSS signals cannot be reliably used, and as a fail-safe alternative, it is possible to improve the reliability of the positioning algorithm. The unmanned aerial vehicle can reach the target position or at least safely land at a pre