The Detection Of Airport Runways Based On Wishart Classifier In PolSAR Image

Polarimetric synthetic aperture radar,referred to as PolSAR,is an advanced imaging radar system operating in the microwave frequency range,and can obtain target information all-day and all-weather.Compared with the conventional unipolar imaging radar,its multi-channel design makes it possible to measure the amplitude and phase of scattering echoes in the combination of various polarization modes,and identify the objects better.Therefore,PolSAR has attracted more and more attention in the many fields,such as image classification,target detection and recognition.As the key part of airport,the automatic detection and recognition of runway plays an important role in many fields,such as autonomous landing system for unmanned aerial vehicle and aircraft,missile precision strike,emergency rescue and so on.In this thesis,a study of airport runway areas detection is introduced based on the scattering characteristics of terrains.The main work are as follows:1.A method for airport runway areas detection based on unsupervised classification is proposed by combining polarized scattering characteristics with runway structure.Firstly,the PolSAR image is roughly classified by the entropy and Freeman decomposition;Then the Wishart classifier is utilized to improve the classification accuracy,and suspected airport areas are extracted from classification result based on the power property of terrain;Finally,structural features are used for final detection of the true runway area.2.In order to solve the problem of large computational complexity,a method for airport runway area detection based on superpixel is proposed in this thesis.Firstly,the simple linear iterative clustering(SLIC)method is utilized to converge the adjacent pixels with similar polarization characteristics in PolSAR image into a size-controllable superpixel.Then superpixels are classified by the polarized scattering characteristics and Wishart classifier.Finally,suspected airport areas are extracted from classification result,and runway structure are applied to identify the true runway area from extracted areas.Experimental results with measured data demonstrate that the two kinds of detection methods can effectively detect the airport runway area with no alarm in the situation of less prior information,and have a good application.And the computational complexity of algorithm 2 is much lower than that ofalgorithm 1.In terms of edge detail preservation and structure integrity,the performance of the algorithm 1 is better than that of the second algorithm.