Anti-interference visual odometer construction method based on optimized Gaussian mixture model

The invention discloses an anti-interference visual odometer construction method based on an optimized Gaussian mixture model, which comprises the steps thatcontinuous video frame sequence images areread and collected, and grayscale processing is performed; similarity determination is conducted on the images, if the similarity is low,then the images are discarded until the judgment result is highly similar; feature points in grayscale images of adjacent frames of afirst n1 frames are extracted; a Gaussian mixture model is established for each feature point; a target detection is conducted byusing the Gaussian mixture model to obtain effective feature points;the effective feature points are taken as initial data to obtain position coordinates of the effective feature points in the currentframe sequence grayscale images and toobtain the effective feature points of subsequent frame grayscale images; effective feature point pairs are composed ofthe obtained effectively feature points between the adjacent grayscal