The invention discloses a grid star map identification method and
system based on weighted voting and geometric
verification, and the method specifically comprises the steps: firstly, constructing a feature mode
database containing all-day navigation
stars, and storing the rasterized geometric mode of an adjacent star
sky region of each navigation star and the inverse document frequency (IDF) weight of each grid unit in the
database; secondly, extracting a plurality of brightest observation
stars as main
stars from an observation star map obtained by a star
map sensor, and generating an observation geometric mode for each main star; inquiring an offline
database by utilizing an observation mode, generating a candidate navigation star
list, dynamically filtering low-credibility candidate stars through an adaptive threshold strategy, and sorting the candidate stars by adopting a mixed scoring function; and finally, carrying out
pose consistency check on the plurality of candidate stars with the highest scores through a geometric
verification method, and identifying the most reliable matching result. On the premise of ensuring the real-time performance, the grid star map recognition robustness and accuracy under various noises and interferences are improved.