The Detection Of Airplane Based On Scattering Characteristics In PolSAR Image

Polarimetric Synthetic Aperture Radar(PolSAR)is an advanced active earth observation device.It has certain penetrability to the vegetative cover,shallow sea and so on.Besides it can achieve day/night and all-weather surveillance for the interested targets.Meanwhile the data collected by PolSAR systems can characterize the polarimetric scattering characteristics of targets exactly.So PolSAR image has been widely used in many fields such as national economy,national defense and so on.Airplane detection is a branch of PolSAR image interpretation.Based on the study of the morphological characteristics and polarimetric scattering characteristics of the airplanes in the PolSAR image,two kinds of airplane detection algorithms are proposed.In the first algorithm,the distinction parameters of the power image and the T22 channel image are calculated respectively based on runway samples first,and distinction parameters are fused.Second,the detection feature is constructed by combining the distinction parameters with feature values extracted from Cloude decomposition.Then,the mean of background clutter is calculated by setting a airplane detection window to traverse the entire image of T22 channel.Finally,threshold segmentation is applied to achieve airplane detection.In order to overcome the dependence of sample information in the former algorithm,the method of airplane target detection based on the Polarization Cross Entropy(PCE)and Yamaguchi decomposition is proposed.First of all,the similarity parameters are calculated between each pixel and three canonical scattering mechanisms,and PCE is redefined based on similarity parameters.Second,Yamaguchi decomposition is adopted to extract double-bounce scattering component power.Then,detection feature is constructed by combining PCE with double-bounce scattering component power.Finally,the technology of threshold segmentation is used to extract airplanes.Multi-look fully PolSAR datasets acquired by U.S.UAVSAR systems and U.S.AIRSAR are adopted to verify the efficiency of the two new algorithms.The experimental results show that the two algorithms can detect the airplanes correctly and effectively under the condition of incomplete morphological features,and the leakage alarms and false alarms are low.Comparing the results of two algorithms,the former algorithm can detect airplanes rapidly,but it needs to select runway samples manually.The latter method is more practical than the former one,because it overcomes the dependence of sample information,but it takes more time than the former algorithm.