Detection Of Airport And Airplane Using PolSAR Images

In recent decades, Synthetic Aperture Radar (SAR) technology has been constantly developed for its advantages in all-day and all-weather rotation, abundant information and the existence of polarmetric information. With the promotion in SAR technology, the application in polarimetric SAR and the emergence of high resolution SAR, its applications in remote sensing are becoming more and more wide and deep, especially in airport ROI detection and related airplane targets. In this paper, the radar image’s polarimetric scattering mechanisms, the targets’scattering characteristics and the information of the images are firstly introduced, then the pre-processing of SAR images is studied on, generally on the methods of image filtering and image enhancement, then the algorithms on airport ROI detection completely and efficiently are studied, finally the ways in which the airplane around the airport are argued.In Chapter1, the basics on SAR data processing and the resources of researched data are the subjects. The processed images stem from two ways-the satellite-operating ALOS’s low resolution sensor, PALSAR and the airborne DLR’s high resolution sensor, ESAR.In Chapter2, the filtering of coherent speckle and the image enhancement technology are two subjects-they belong to the preprocessing technology. In speckle filtering, the multiplicative noise model based algorithms are focused on. Refined Lee-sigma filtering, for instance, is the latest algorithm in this field. The ALOS data is filtered, and ENLs of both pre-filtering and filtered images are compared. In this way, the usefulness of the algorithm is proved. In the image enhancement, the concept of Shannon entropy of SAR data is introduced, and its contribution in the enhancement of object form the background is proved.In Chapter3,3approaches of detection of airport ROI in SAR images are studied. They are namely the one based on CV model, the one based on the maxim fuzzy entropy and the one based on ROEWA detector. The former two are based on region detection, as the latter one is based on edge detection. The algorithm based on CV model makes use of the speckle filtering and SE1parameter to improve the amplitude of the edge and plays well in the detection of airport ROI in high resolution SAR images. With different parameters, the ROIs in different sizes are obtained. The algorithm based on the maxim fuzzy entropy takes the advantage of the ABC algorithm’s positive feedbacks, parallel computing and robustness, and evolve the efficiency of finding the accurate threshold. The algorithm based on ROEWA detector is born form the multi-edge model, and it proves to maintain the constant false rate in the edge detection in SAR images.In Chapter4, the algorithms on airplane detection in SAR image are studied, including the CFAR algorithms and the algorithm based on fractal features. The different CFAR algorithms, such as CA-CFAR, OS-CFAR, SO-CFAR, GO-CFAR and VI-CFAR are employed in images and compared. In high resolution image, the algorithm based on fractal features is used and the weakness of algorithms above is argued.