The Research And Application Of Object Detection And Recognition Algorithm In Airport Surface Based On Computer Vision

Computer vision surveillance technology is drawing more and more attention, and airport surface surveillance is the most important need in the area of civil aviation, computer vision surveillance will become an integral part of the airport surface surveillance system. So study on the target detection and recognition in the airport surface is meaningful.In this paper, computer based image processing technology used into research the detection and recognition of surface. First, an overview of the current research and development situation of computer image detection and recognition. Based on the conclusion, make the definition of the constraint conditions of airport surface detection and extraction of feature. SIFT, SURF, PCA-Sift, ASIFT were analyzed by experimental simulation and performance comparison and get a suitable algorithm for airport target detection.In the next place, establish and make the improvement of the BOW model in this paper, put forward a kind of pyramid BOW model that based on the core area in a image. In this model, the core area in the image will be detected and descript only.Finally, the paper put forward a method of multi target detection and recognition in surface, and it is based on geometry feature and gray feature. According to the characteristics of ground target, the fusion of geometry feature like rectangular and circular degree and Invariant moments are used to complete the matching and recognition of target in surface. A recognition system software is completed to do the detection and identification of aircraft, car, people. And based on the proposed algorithm and model, two different classification and recognition experiment system were designed. Through the comparison of analysis results verified the validity of the improved model and efficiency of the algorithm.