The traffic environment is becoming more and more complex

and the application scenarios of vehicle identification technology are becoming more and more extensive. In this paper

the traditional orb algorithm is improved by clustering analysis method and applied to the front vehicle recognition system. The feature vectors extracted from ORB are classified and filtered to reduce the number of noise points and feature redundancy points

so as to improve the matching efficiency. The experimental results show that the orb algorithm based on K-means eliminates the interference of irrelevant objects to a certain extent. It not only retains the advantages of fast operation of the original ORB algorithm

but also improves the recognition accuracy of common databases by about10 percentage points

compared with similar algorithms The improved algorithm has better application effect and applicability in the field of intelligent transportation in complex environment.