Traditional jamming methods protect scene targets by generating barrage jamming or deceptive jamming effect, but the single jamming effect is currently difficult to meet the increasingly complex electromagnetic environment. To meet the complex electromagnetic environment, a 2-D precise controllable smart jamming method is proposed in this letter, which can produce two kinds of jamming effects: barrage jamming and deceptive jamming. Based on the synthetic aperture radar (SAR) imaging properties of linear frequency modulation (LFM), the range and azimuth modulation terms have been designed to generate precise controllable jamming. First, both the range and azimuth jamming positions can be controlled by first-order phase error modulation. Subsequently, multiple phases sectionalized modulation is performed in the range direction, quadratic phase error or periodic phase error modulation is performed in the azimuth direction. Different jamming effects can be generated by changing the jamming modulation item. Theoretical analysis and simulation results show that the proposed method can produce 2-D precise controllable smart jamming effects.

Graph attention network can generate effective feature embedding by specifying different weights to different nodes. The key of the research on heterogeneous graph embedding is the way to combine its rich structural information with semantic relations to aggregate the neighborhood information. Most of the existing heterogeneous graph representation learning methods guide the selection of neighbors by defining various meta-paths on heterogeneous graphs. However, these models only consider the information contained in the nodes under different paths and ignore the potential semantic relationships of nodes in different neighbor graph structures, which leads to the underutilization of graph structure information. In this article, we propose a novel adaptive framework named Neighbor Graph Aggregated Graph Attention Network (NGGAN) to fully exploit graph topological details in heterogeneous graph, and aggregates their information to obtain an effective embedding. The key idea is to use different levels of sampling methods to define neighborhood, and use neighbor graphs to represent the complex structural interaction between nodes. In this way, the high-order relationship between nodes and the latent semantics of neighbor graphs can be fully explored. Afterward, a hierarchical attention mechanism is applied to adaptively learn the importance of different objects, including node information, path information, and neighbor graph information. Multiple downstream tasks are performed on four real-world heterogeneous graph datasets, and the experimental results demonstrate the effectiveness of NGGAN.

Soil moisture is an important parameter affecting environmental processes such as hydrology, ecology and climate. Microwave remote sensing is an effective means of surface soil moisture measurement. Aiming at the influence of vegetation cover in the process of surface soil moisture inversion of winter wheat farmland by microwave remote sensing, a cooperative inversion method using multi-source remote sensing data is proposed in this paper. Thirty-three feature parameters are extracted from Radarsat-2 full polarization SAR data and Sentinel-2 optical data, and ten parameters with high correlation with soil moisture are selected to participate in soil moisture inversion by Pearson correlation analysis. Combined with the ground sampling data, four machine learning models, including Random Forest, Generalized Regression Neural Network, Radial Basis Function and Extreme Learning Machine, are used for quantitative inversion of soil moisture to reduce the impact of vegetation and improve the inversion accuracy. The experimental results show that the Random Forest model is the optimal. The average of determination coefficient is 0.63959, and the average of root mean square error is 0.0317 cm3/ cm3, which provides a reference for the inversion of soil moisture in farmland using multi-source remote sensing data.

The concept of spaceborne interrupted frequency-modulated continuous-wave (IFMCW) synthetic aperture radar (SAR) subverts the design principles of the traditional frequency-modulated continuous-wave (FMCW) SAR system, which uses a single physical antenna to transmit signal and receive echo alternatively. Due to the demand of turning off the receiver in the process of transmitting signal, periodical data-missing phenomena will be aroused in the echo data, and the missing ratio is close to 50%. When traditional imaging algorithms were used to process the echo data, the artifacts, appearing as periodically replicated false targets, will be generated in the focused SAR image due to the periodical missing data, thus affecting the image interpretation. In order to effectively suppress the artifacts, a novel method exploiting singular value threshold-based matrix completion (SVT-MC) technology was proposed in this letter, where the singular value decomposition of the Hankel matrix was utilized to recover the missing data with high precision. Finally, the experimental results based on both point targets and distributed targets verify the effectiveness and superiority of the proposed method.

Currently, the spaceborne synthetic aperture radar (SAR) system transmits a great deal of data to the ground processing station and generates massive images daily, only a tiny fraction of which contains radio frequency interference (RFI). However, most of the existing RFI detection methods are based on the prior conditions in which the known image contains interference. In fact, it is difficult to learn whether SAR images contain RFIs without prescreening, so it is of great significance to the rapid and real-time screening and detection of RFI in SAR images. This paper proposes a method to screen and detect RFI from massive SAR images simultaneously. 1) We construct an approximate RFI-free background image by using the preprocessed time-series SAR images acquired in the past. 2) We generate difference images based on the image change detection method and analyze them by using an adaptive threshold, then calculate the entropy of all difference images to complete the preliminary screening of the RFI-containing images. 3) According to the preliminary results, we remove the RFI-containing parts; after reconstructing the background, we repeat step 2 with the images to be detected and obtain the final screening and detection results. Massive experimental results based on Sentinel-1 images validate the performance of the proposed method.

In order to realize the raw data simulation of high resolution squint spotlight Synthetic Aperture Radar (SAR) efficiently, an effective acceleration algorithm is proposed. This algorithm combines the time-domain raw data simulation model and its signal characteristics to compensate the range cell migration (RCM) existing in the raw data simulation process of squint spotlight SAR. An adaptive data partitioning algorithm is used, and computes partitioned data respectively in GPU. Then the partitioned data are transmitted and spliced. The algorithm improves the computational efficiency of time-domain raw data simulation, and it solves the problems of huge volume of raw data, limitation of GPU memory and data transmission. The experimental results of point target and distributed target show that the speedup ratio of this algorithm reaches 209.93, which verifies the effectiveness of the proposed method.