Mosaic mode synthetic aperture radar (SAR) combines the spotlight or sliding spotlight mode with a ScanSAR mode to accomplish the high azimuth resolution and large imaging swath SAR imaging. This paper studies the problem of the frequency modulated continuous wave (FMCW) SAR imaging with a Mosaic mode, where the imaging algorithms for pulse radars are not suitable for a FMCW SAR. To deal with this issue, this paper proposes a FMCW Mosaic mode SAR imaging algorithm based on modified polar formation algorithm (PFA), which considers the radar motion during the signal transmission and performs the SAR imaging operation for spotlight mode in each Mosaic unit. The simulation results of point targets verify the effectiveness and feasibility of the proposed method, which may provide a valuable reference for the application of Mosaic system to miniaturized platforms.

Using deep convolutional neural networks to achieve automatic target recognition (ATR) is effective but it will bring heavy computation burden. Here we propose an optronic convolutional neural network (OPCNN) to realize ATR in optics. By using OPCNN, the computation cost is dramatically reduced and good performance of recognition accuracy is obtained simultaneously. Experimental results on Moving and Stationary Target Acquisition and Recognition dataset demonstrate the feasibility of our proposed OPCNN architecture.

The azimuth multichannel spaceborne SAR system can achieve the high-resolution and wide-swath simultaneously. In a high-resolution and wide swath (HRWS) SAR system, the Doppler centroid (DC) frequency estimation is an important step for Doppler ambiguity signal reconstruction, which should be precisely estimated. In this paper, we propose a novel Doppler centroid frequency estimation method for HRWS-SAR system based on modified shuffled frog leafing algorithm (MSFLA). Firstly, the cost function is designed based on the average power difference between ambiguities and useful signal. Then, the MLSFA algorithm is proposed to accomplish the DC estimation. Finally, a good HRWS-SAR image is obtained based on the estimated DC. Compared with the traditional time-domain based DC estimation methods, the proposed method is not affected by the troubled cross-terms induced by high-order moments. The effectiveness of the proposed method is verified by the simulation experiments and real-measured SAR data.

Radar forward-looking imaging has important application value in target attack and terrain detection. However, traditional synthetic aperture radar (SAR) cannot perform high-resolution imaging of the area directly in front of the flight trajectory, due to the limitation of mechanism. This paper proposes a new airborne forward-looking imaging scheme based on maneuvering motion. The scheme firstly starts from the movement trajectory of the platform, gives the spatial geometric relationship of the forward -looking imaging, and establishes the echo signal model, further analyzes the azimuth resolution characteristics of the signal. On this basis, combined with the phase characteristics of the echo signal, a forward-looking imaging algorithm suitable for this scheme is derived. Finally, through simulation experiments, the target imaging results are analyzed to verify the effectiveness of the algorithm.

Multichannel Terrain Observation by Progressive Scans (TOPS) SAR has overcomed the system-inherent limitation of conventional synthetic aperture radar (SAR). Multichannel TOPS SAR is capable of imaging a scene with high geometric resolution and wide swath. But the complexity of algorithms to process multichannel TOPS raw data has increased. Multichannel TOPS takes much more time and computer resource to focus an image than conventional SAR does. On the other hand, optical computing has the advantages of high speed, huge capacity and low power. So we tried to use optical computing, especially optical two dimensional Fourier transform, to process multichannel TOPS raw data in this paper. Due to the limitation of system devices, we just partly realized the computing on a 4-f system. In our method, the raw data was firstly compensated to a signal which could be seen as the raw data of another Stripmap system without range cell migration. In the 4-f system, the Stripmap signal is then transformed into its two dimensional frequency domain by a lens and multiplied by a filter to focus the image. We had tested our system with simulation point target signals. The result showed that the proposed method could quickly process Multichannel TOPS data with resolution loss smaller than 11%.

Narrow-band interference (NBI) is a major threat for synthetic aperture radar (SAR) system, which degrades the imaging quality severely. In this paper, a new type of narrow-band interference (NBI) is investigated, i.e., periodicity transmitted NBI (PT -NBI). Compared with the conventional NBI, PT -NBI is intermittently emerged in one azimuth pulse, which shows periodic property in the range time domain and broadband characteristic in the range frequency domain, resulting in the severe performance degradation by using the existing NBI suppression methods. The distinctions and characteristics between the NBI and PT -NBI in different transform domains are analyzed firstly. On this basis, a PT -NBI suppression method is proposed by applying the joint time domain detection and Eigensubspace (ESP) filtering in this paper, which can effectively realize the interference suppression for PT - NBI.

Ground Penetrating Radar (GPR) is a non-destructive technique that employs electromagnetic waves to map subsurface structures. An effective simulation method is essential for GPR system design and signal processing. In this paper, a simulation method of GPR is proposed for dispersive media that is usually encountered in GPR applications. The behavior of dispersive media is simulated in our method by filtering the transmitted signal of GPR with an approximation of the system transfer function that is obtained by interpolating the values of the system transfer function evaluated at different frequencies, each of which is acquired by simulating GPR using the existing ray tracing based method of GPR simulation for non-dispersive media with a mono-frequency signal of the corresponding frequency as its transmitted signal. The accuracy and efficiency of our method were validated by comparing the experimental results of our method to those of the Frequency-Dependent Finite-Difference Time-Domain (FD-FDTD) method. The experimental results demonstrate that our method can simulate GPR for dispersive media with acceptable accuracy consuming much less running time than the FD-FDTD method for GPR simulation for dispersive media.