Unmanned aerial vehicle image splicing method based on vectorization feature screening

Abstract

The invention discloses an unmanned aerial vehicle image splicing method based on vectorization feature screening, and the method comprises the steps: extracting feature points of each image, and obtaining feature point matching pairs; constructing the matched feature point pairs into matched vectorized feature pairs; randomly selecting a vectorized feature pair, and calculating a first similarity transformation matrix parameter; counting the number of the vectorized feature pairs conforming to the first similarity transformation matrix, and obtaining a set with the maximum number of the vectorized feature pairs after iteration; recovering the vectorized feature pairs in the vectorized feature pair set with the largest number to obtain feature point pairs, randomly selecting the feature point pairs, calculating parameters of a second similar transformation matrix, counting the number of the feature point pairs conforming to the second similar transformation matrix to obtain a final transformation matrix, and obtaining the relative position of the image according to the transformation matrix obtained through calculation. And a final spliced image is obtained through the image fusion step. According to the method, the feature screening accuracy can be effectively improved, a good splicing effect is obtained, and the actual industrial application requirements are met.

Images