Fan Jie，Si Weisen，Ba Xiang
China Southern Airlines Henan Branch，Zhengzhou 450000，China
Abstract: In order to identify the airframe damage region clearly and quickly, the K-means clustering algorithm is introduced to process the damage image. By analyzing the limitations of the K-means clustering algorithm, an improved method of clustering iteration termination condition based on error square sum difference and pixel change is proposed. Through the verification of aircraft body damage images, it is verified from two aspects: damage recognition effect and operation efficiency. The verification results show that the airframe damage identification method based on improved clustering not only ensures the identification effect of image damage regions, but also significantly reduces the number of iterations, improves the operation efficiency of the clustering algorithm, and can meet the requirements of efficient processing of airframe damage identification.
Key Words: damage region division; image recognition; K-means clustering algorithm