The Research On Land Target Detection Method For Airborne SAR Images

Synthetic Aperture Radar (SAR) has been widely used in the field of aerospace, earthobservation and national defense, with the development of unmanned aerial vehicle recentyears, target detection technology based on airborne SAR images has been one of the coretechniques of SAR military applications. The ground tanks group, transport vehicles andairports, are all important detection targets in the field for military, when the automatic andreal-time detection is applied to these targets, accurate intelligence information can beobtained, which play an important role for battle reconnaissance and military command.The main work and fruit of this dissertation is listed as follows:(1) The current research status at home and abroad of target detection of SAR imagesis analyzed and the main work of this dissertation is described. The characteristics of SARimages are analyzed. For the ground target detection and the vehicles and airports are takenas examples, the detection features are analyzed and the corresponding detection schemesare designed. A fast filtering algorithm based on edge detection is analyzed and improved inthis dissertation, compared with the classical filtering algorithms, the test resultsdemonstrate that the capability of edge preserving is improved obviously and the noise issuppressed efficiently.(2) Vehicle target detection algorithms based on constant false alarm rate (CFAR)technology are studied. The main background clutter models are introduced and analyzed,at the same time, the database of vehicle background clutter models is established. GlobalCFAR algorithm, double parameter CFAR and adaptive CFAR based on the statisticalcharacteristics of local clutters are studied, and the adaptive CFAR algorithm that use of thedatabase is studied and which is used to vehicle detection. The result of experiments showsthe algorithm is improved in detection rate and calculation efficiency.(3) Airport target detection algorithms based on feature segmentation are studied. TheSeveral representative segmentation algorithms of SAR images are analyzed. Adouble-thresholds segmentation algorithm based on gray contrast is improved to detectairport and a composite identification method is adopted to exclude false targets. The resultof experiments shows the target can be detected more accurately and quickly by theimproved.