Improved Small Target Detection Method for SAR Image Based on YOLOv7

In order to solve the problems that the current synthetic aperture radar(SAR)image target detection method cannot adapt to targets of different sizes,and the complex image background leads to low detection accuracy,an improved SAR image small target detection method based on YOLOv7 was proposed in this study.The proposed method improved the feature extraction network by using Switchable Around Convolution(SAConv)in the backbone network to help the model capture target information at different scales,thus improving the feature extraction ability for small targets.Based on the attention mechanism,the DyHead module was embedded in the target detection head to reduce the impact of complex background,and better focus on the small targets.In addition,the NWD loss function was introduced and combined with CIoU loss.Compared to the CIoU loss function typically used in YOLOv7,the NWD loss function pays more attention to the processing of small targets,so as to further improve the detection ability of small targets.The experimental results on the HRSID dataset indicate that the proposed method achieved mAP@0.5 and mAP@0.95 scores of 93.5%and 71.5%,respectively.Compared to the baseline model,this represents an increase of 7.2%and 7.6%,respectively.The proposed method can effectively complete the task of SAR image small target detection.

基于小波域隐Markov模型的SAR图像滤波方法

在小波域隐Markov模型(HMM)的基础上提出一种新的合成孔径雷达(SAR)图像的滤波方法。首先根据小波变换的内在特征,建立小波域的隐Markov树(HMT)模型,通过EM算法可以获得该HMT模型的参数估计。然后根据SAR图像的统计性质,将SAR图像的乘法斑点杂噪声在局部范围内近似为加性白高斯噪声,通过最小均方差(MMSE)估计可以获得信号的小波变换值。通过对真实SAR图像的应用,结果说明该方法可以在保存图像细节特征的情况下有效地抑制图像的噪声。

Advance of Rice Recognition and Monitoring by SAR

Synthetic Aperture Radar（SAR） is a more effective remote sensing data source for rice recognition and monitoring than optical remote sensing in the regions with more wet and cloudy sky due to its all-weather, all-time, high resolution and wide covering characteristics. This paper summarizes SAR types and their feature used for rice study, introduces the backscattering model for rice monitoring, and analyses the main factors influencing backscattering coefficient. The studies of rice recognition and monitoring based on SAR in domestic and abroad are reviewed and the futures in the related areas are prospected.

PHASE AUTOFOCUSING ALGORITHM FOR COMPRESSED INVERSE SYNTHETIC APERTURE RADAR IMAGING

As same as the conventional inverse synthetic aperture radar（ISAR）, the compressed ISAR also requires the echo signal based motion compensation, which consists of the range alignment and the phase autofoeusing. A phase autofocusing algorithm for compressed ISAR imaging is presented. In the algorithm, phase autofocusing for the sparse ISAR echoes is accomplished using the eigenvector method. Experimental results validate the effectiveness of the algorithm.

SAR Moving Target Detection and Imaging Based on WVH Transform

Aim To propose a generalized and closed representation of the Wigner Ville Hough transform(WVHT), for the moving target detection and imaging in the design of synthetic aperture radar(SAR). Methods Based on the line integral, the WVH transform was derived by combining the Wigner Ville distribution (WVD) and the Hough transform (HT) together. The new transform was then verified with computer by the simulated SAR echoes. Results and Conclusion The correctness and the validity of the WVH transform were proved by the computer simulation. Compared with the conventional WVD HT method, the new approach based on the WVHT can simplify the processing procedure, it can translate the chirp echoes of multi targets of SAR from the time domain into the parameter space directly, while suppressing the cross terms of WVD and estimating the motion coefficients for the final imaging. It is obvious that the WVH transform can be also used in other cases for the chirp signal detection.

SAR IMAGE RECOGNITION BASED ON MULTI-ASPECT OF SHADOW INFORMATION

The traditional synthetic aperture radar（SAR） image recognition techniques focus on the electro magnetic （EM） scattering centers, ignoring the important role of the shadow information on the SAR image recognition. It is difficult to classify targets by the shadow information independently, because the shadow shape is dependent on the radar aspect angle, the depression angle and the resolution. Moreover, the shadow shapes of different targets are similar. When the multiple SAR images of one target from different aspects are available, the performance of the target recognition can be improved. Aimed at the problem, a multi-aspect SAR image recognition technique based on the shadow information is developed. It extracts shadow profiles from SAR images, and takes chain codes as the feature vectors of targets. Then, feature vectors on multiple aspects of the same target are combined with feature sequences, and the hidden Markov model （HMM） is applied to the feature sequences for the target recognition. The simulation result shows the effectiveness of the method.

ISAR MOTION COMPENSATION USING ROPE

Motion compensation is a key step for inverse synthetic aperture radar (ISAR) imaging. Many algorithms have been proposed. The rank one phase estimation (ROPE) algorithm is a good estimator for phase error widely used in SAR. The ROPE algorithm is used in ISAR phase compensation and the concrete implementation steps are presented. Subsequently, the performance of ROPE is analyzed. For ISAR data that fit the ROPE algorithm model, an excellent compensation effect can be obtained with high computation efficiency. Finally, ISAR real data are processed with ROPE and its imaging result is compared with that obtained by the modified Doppler centroid tracking (MDCT) method, which is a robust and good estimator in ISAR phase compensation.

Comparison on Winter Wheat Yield Estimating Models Based on Radarsat-2 and HJ Satellite in Huaihe River Region

The establishment of crop yield estimating model based on microwave and optical satellite images can conduct the mutual verification of the accuracy of the reported crop yield and the precision of the estimating model. With Shou County and Huaiyuan County of Anhui Province as the experimental fields of winter wheat producing areas, the linear winter wheat yield estimating models were established by adopting backscattering coefficient and Normalized Difference Vegetation Index（NDVI） based on images from the synthetic aperture radar（SAR）—RDARSAT-2 and HJ satellite photographed in mid-April and early May, 2014, and then comparisons were conducted on the accuracy of the yield estimating models. The accuracies of the yield estimating models established using co-polarized（HH） and cross-polarized（HV） modes of SAR in Jiangou Town, Shou County were 68.37% and 74.01%, respectively, while the accuracies in Longkang Town, Huaiyuan County were 63.10%and 69.10%, respectively. Accuracies of yield estimating models established by HJ satellite data were 69.52% and 66.43% in Shou County and Huaiyuan County, respectively. Accuracies of winter yield estimating model based on HJ satellite data and that based on SAR were closed, and the yield difference of winter wheat in the lodging region was analyzed in detail. The model results laid the foundation and accumulated experience for the verification, parameters correction and promotion of the winter wheat yield estimating model.

ENLARGING SAR PATCH-MAPPING AREA BY ANTENNA BEAM SCAN

The approach for enlargement of SAR patch mapping area by antenna beam scan is investigated, which serves for moderate fine-resolution mapping of medium-sized terrain patches. The scanning angular velocity and the scanning angular scope are determined respectively. The angular velocity of the scanning antenna is controlled to scan over just one azimuth 3 dB beam width in the time interval during which the radar platform moves over one synthetic aperture length determined from the desired cross-range resolution, radar wavelength, nominal slant range, and squint angle. The scanning angular scope is mainly determined by the azimuth width of the terrain patch, nominal slant range, squint angle, platform velocity, and azimuth beam width. Finally, the related experimental results of an airborne SAR are presented. The linear range-Doppler algorithm is employed in image formation after motion compensation is conducted to remove the effect of transnational motion of the radar platform relative to the map center.

DEVELOPMENT OF MOVING TARGET DETECTION AND IMAGING BY AIRBORNE SAR

The detection and ima ging of moving targets based on airborne synthetic aperture radar (SAR) is a cru cial technique for the modern radar. Firstly, the mathematical model of SAR ech o signal which comes from moving targets is constructed. Based on this model, th e features of moving target imaging are introduced and the effects of target mov ement to SAR imaging are analyzed. Then the development and the status of this t echnique are reviewed in detail. Finally, some frontiers of this field are point ed out.

Suppression of Speckle in SAR Images Using Wavelet-Based HMM

In order to suppress the speckle appearing in synthesis aperture radar (SAR) images, a novel speckle reduction method based on wavelet domain hidden Markov tree (HMT) was proposed. First, the image was logarithmic transformed to change the statistical property of the speckles. Then an HMT was constructed in the correspondent wavelet domain. Based on this model, the image signal was restored by maximum likelihood estimation and speckle noise was suppressed. Simulating SAR images had shown that the performance of the filter is satisfactory for both speckle smoothing and edges presentation, and for generating visually natural images as well.

SAR MOTION INFORMATION SENSOR BASED ON GNSS/SINS

Synthetic aperture radar (SAR) is theoretically based on uniform rectilinear motion. But in real situations, the flight cannot be kept in a uniform rectilinear motion due to many factors. Therefore, the motion compensation is needed to achieve the high-resolution image. This paper proposes an improved motion information sensor (MIS)-based on global navigation statellite system (GNSS) and strapdown inertial navigation system (SINS) for SAR motion compensation. MIS can provide the long-term absolute accuracy, and the short-term high relative accuracy during SAR imaging. Many issues related to MIS, such as system design, error models and navigation algorithms, are stressed. Experimental results show that MIS can provide accurate navigation information (position, velocity and attitude) to meet the requirements of SAR motion compensation. Especially, MIS is suitable for the case: the accuracy of airplane master inertial navigation system is too low or not configured.

主星带伴随分布式小卫星SAR系统效能分析

针对主星带伴随分布式小卫星系统效能分析中几个关键问题进行了研究。首先,根据分布式小卫星SAR 系统的任务需求,给出了系统效能评估指标体系;结合分布式小卫星轨道运行特点,建立了系统能力模型;根据系统的工作状态,建立了系统可靠性Markov模型;基于能力模型与可靠性模型,建立了主星带伴随分布式小卫星SAR 系统效能评估模型。仿真结果证明,以上模型的建立为主星带伴随分布式小卫星SAR系统效能分析以及系统参数设计提供理论依据。

Entropy function optimization for radar imaging

The convergence performance of the minimum entropy auto-focusing（MEA） algorithm for inverse synthetic aperture radar（ISAR） imaging is analyzed by simulation. The results show that a local optimal solution problem exists in the MEA algorithm. The cost function of the MEA algorithm is not a downward-convex function of multidimensional phases to be compensated. Only when the initial values of the compensated phases are chosen to be near the global minimal point of the entropy function, the MEA algorithm can converge to a global optimal solution. To study the optimal solution problem of the MEA algorithm, a new scheme of entropy function optimization for radar imaging is presented. First, the initial values of the compensated phases are estimated by using the modified Doppler centroid tracking （DCT）algorithm. Since these values are obtained according to the maximum likelihood （ML） principle, the initial phases can be located near the optimal solution values. Then, a fast MEA algorithm is used for the local searching process and the global optimal solution can be obtained. The simulation results show that this scheme can realize the global optimization of the MEA algorithm and can avoid the selection and adjustment of parameters such as iteration step lengths, threshold values, etc.

ISAR MOTION COMPENSATION USING MODIFIED DOPPLER CENTROID TRACKING METHOD

With regard to the phase compensation in inverse synthetic aperture radar (ISAR),the modified Doppler centroid tracking (MDCT) method is developed which applies the phase gradient autofocus (PGA) algorithm developed by Wahl[1]to improve the Doppler centroid tracking (DCT) method[2].When the phase compensation is performed,the proposed approach smartly eliminates the effect of the rotational phase component (RPC) on the estimation of the translational phase component (TPC) by circular shifting,windowing and iteration steps. After several iterations,the maximum likelihood estimation and compensation of the TPC of the target can be realized more effectively.The processing results of live data show that the proposed method can improve the imaging quality of ISAR significantly.

FAST IMPLEMENTATION OF CONVOLUTION BACKPROJECTION ALGORITHM IN SPOTLIGHT SAR

A fast implementation of the convolution backprojection（CBP）algorithm in spotlight synthetic aperture radar（SAR）is presented based on the fast Fourier transform（FFT）.Traditionally,the computation of the ＇backprojection＇ process is expensive,since resampling in the process is implemented by using the interpolation operation.By analyzing the relative location relationship among different pixels,the algorithm realizes the ＇backprojection＇ using a series of FFTs instead of the interpolation operation.The point target simulation validates that the new algorithm accelerates the CBP algorithm,and the computational rate increases about 85%.

AN EDGE DETECTOR BA SED ON WAVELET TRANSFORM FOR SAR IMAGE

Edge detecting is a f undamental issue in image analysis and image processing. In this paper, based on wavelet transform, an edge detector of SAR image is studied to serve the image aided navigation. Though wavelet transform is efficient as an edge detector, a l ot of false edges will be extracted from a SAR image, especially in areas of hig h reflectivity, with wavelet transform and a fixed threshold because SAR images are always contaminated by speckle. Therefore, this paper develops an algorithm by using a variable threshold with constant false alarm rates to help suppressin g the speckle in SAR images. Finally, two simulating tests are given and the res ults show that the effect of the multiplicative speckle is controlled and there are few false edges in the gotten edge images. Moreover, owing to the threshold from the wavelet transform coefficients, the algorithm can assure the real time of SAR image aided navigation.

A novel method for determining the anisotropy of geophysical parameters: unit range variation increment(URVI)

Geometric anisotropy is commonly assumed in the investigation of the spatial variations of geophysical parameters. However, this assumption is not always satisfied in practice. We propose a novel method to determine the anisotropy of geophysical parameters. In the proposed method, the variograms are first normalized in all directions. Then, the normalized samples are fitted by the unit range variation increment（URVI） function to estimate the intensities of the variograms in each direction, from which the anisotropy can be finally determined. The performance of the proposed method is validated using InSAR atmospheric delay measurements over the Shanghai region. The results show that the deviation of the method is 6.4%, and that of the geometric anisotropy-based method is 21.2%. In addition, the computational efficiency of the new method is much higher. Subsequently, the URVI- and the geometric anisotropy-based methods are cross-validated in the cross-validation experiments by using Kriging interpolation. The results demonstrate that the structure functions generated with the proposed method are more accurate and can better refl ect the spatial characteristics of the random fi eld. Therefore, the proposed method, which is more accurate and effi cient to determine the anisotropy than the conventional geometry anisotropy-based method, provides a better foundation to estimate the geophysical parameters of interest.

Spatial-temporal Characteristics of Land Subsidence Corresponding to Dynamic Groundwater Funnel in Beijing Municipality,China

Due to long-term over-exploitation of groundwater in Beijing Municipality, regional groundwater funnels have formed and land subsidence has been induced. By combining a groundwater monitoring network, GPS monitor- ing network data, radar satellite SAR data, GIS and other new technologies, a coupled process model based on the dy- namic variation of groundwater and the deformation response of land subsidence has been established. The dynamic variation of groundwater fimnels and the land subsidence response process were analyzed systematically in Beijing. Study results indicate that current groundwater funnel areas are distributed mainly in the southwest of Shunyi District, the northeast of Chaoyang District and the northwest of Tongzhou District, with an average decline rate of groundwa- ter level of 2.66 rn/yr and a maximum of 3.82 m/yr in the center of the funnels. Seasonal and interannual differences exist in the response model of land subsidence to groundwater funnels with uneven spatial and temporal distribution, where the maximum land subsidence rate was about --41.08 mm/yr and the area with a subsidence rate greater than 30 mm/yr was about 1637.29 km2. Although a consistency was revealed to exist between a groundwater funnel and the spatial distribution characteristics of the corresponding land subsidence funnel, this consistency was not perfect. The results showed that the response model of land subsidence to the dynamic variation of groundwater was more revealing when combining conventional technologies with InSAR, GIS, GPS, providing a new strategy for environmental and hydrogeological research and a scientific basis for regional land subsidence control.

Extended Chirp Scaling Algorithm for Spotlight SAR

Among Synthetic Aperture Radar(SAR)image formation algorithms,the chirp scaling algorithm(CSA)is computationally highly efficient and has been practically used.In this paper,how to implement the subaperture extended CSA in a high resolution spotlight mode SAR is studied in detail.The procedures such as subaperture processing and azimuth processing are discussed.According to the problems presented when using the method(A.Moreira,et al,1996)to process spotlight SAR data,full formula expressions suitable for high squinted angles is presented.While point target simulation is analyzed,the method is validated by E SAR raw data.The results can be directly implemented in the design of SAR processors.