Research On The Detection And Recognition Algorithms Of Targets On The Airport Surface

As China's rapid development of the civil aviation industry, the flow of the aircraft increases constantly, the activities on the airport surface become more and more complicated, thus, the efficiency of the ATC system and capability of the surface monitoring need to be improved urgently. Based on Computer Vision Theory, the detection and recognition algorithms of aircraft on the airport surface are applied to the airport surveillance fields. This paper is to detect the generalized multi-scale characteristics(such as grayscale, relative location, etc.) according to the Vision Measurement Principle, finally to achieve the detection and recognition function of the aircraft on the airport surface.In this paper, the detection algorithm of the aircraft target in the airport image was researched, and finally the recognition of the aircraft target parked on the airport with high accuracy was realized.Firstly, the general image of the static object recognition processes was introduced, which was mainly divided into pre-processing, image segmentation,morphological processing and targeting, and each of these steps was discussed.Secondly, the detection of the aircraft based on the edge-detection operator was studied. We mainly discussed four commonly used edge-detection operators which are Roberts operator, Prewitt operator, Sobel operator and Canny operator, and their calculation principles. Then according to the simulation results of the experiment, the performance of the four kinds of operators was compared, and also, the combination with different pre-processing methods was analyzed.Then, the paper discussed the recognition method based on threshold. Otsu and bimodal method were improved on combination with the pre-processing and the threshold choosing problem, and the combination has been realized according to the merits of the two methods. After the test using surveillance pictures taken in Tianjin Binhai International Airport, simulation results showed that the method based on image recognition threshold can recognize the target effectively, and mark thecentroid position of the aircraft.Finally, a template method for target identification was used, the advantages and disadvantages of this method were discussed based on the simulation results.Some achievements have been gained by several methods, the detection results meted the requirements of precision on airport surface.