Study On Bistatic SAR Imaging Algorithms

Bistatic synthetic aperture radar(BiSAR)is one special type of SAR.Synthetic aperture radar(SAR)can be placed on many platforms,such as aircraft,satellite and so on,to detect and image the target detection area.It is of long-distance,all-weather,high-resolution imaging.The transmission signal used in the SAR system has certain ability of penetration,which can penetrate the clouds,rain,fog and even vegetation in order to detect and image the target area.The transmitter and receiver of BiSAR can be loaded on different platforms,whose special configuration improves the flexibility of data collection,making the application scope wider and more flexible.Meanwhile,its special configuration also enables the BiSAR to obtain certain information of non-backscatter.Based on the above advantages,BiSAR has been widely developed in recent years.With the development of science and technology,SAR imaging technology and image processing technology are required to be more efficient and reliable in both battlefield information detection and geological disaster detection application scenarios.The research on BiSAR imaging technology is also very important and necessary.The range equation of BiSAR is expressed mathematically by the addition of two square–root terms.And the principle of stationary phase(POSP)used in monostatic SAR cannot be used to obtain the accurate expression of the 2-D spectrum of BiSAR.Also,the traditional monostatic SAR imaging algorithms can not be directly applied to the BiSAR imaging.In order to solve the above problems,this paper mainly studies on two aspects:(1)the 2-D spectrum of BiSAR;(2)the imaging algorithms of BiSAR.The main content of this paper can be summarized as follows:In the first part,a kind of equivalent wavelength along the direction of beam center is proposed,and based on the physical and geometric relationships,The 2-D spectrum of BiSAR and an Omega-K algorithm are derived.The derivation is under the condition of azimuth time synchronization.The method has avoided using POSP and obtains high accuracy.And then,based on the 2-D spectrum,the specific compensation function of BiSAR Omega-k algorithm is calculated.Experimental results show that the algorithm works well.In the second part,BiSARNet based on the complex convolution and fully convolution is proposed.The complex convolution is applied to the complex signal processing of SAR to extract the complex domain features of SAR,which are further intergrated by real multichannel convolution to obtain the real domain features of SAR.Finally,the real domain features are expanded by using the deconvolution layer and real convolution is used to decode the features and get the final image.BiSARNet uses fully convolution and can accept input with any size.The experiment uses 256 × 256 input in the training process,and can directly input 1024 × 1024 or even larger data in the test.The input and output size is flexible.BiSARNet has fast imaging speed which plays a great role in practical application.The imaging results of BiSARNet can completely present the geomorphic informations,but there are some deficiencies in imaging small targets that is slightly fuzzy,which can not temporarily meet the application requirements of high precision.