Impact crater detection based on remote sensing images (RSI) plays a key role in deep space exploration. However, remote sensing images of planetary surfaces often suffer from low resolution, making the detection of smaller craters a challenging task. To address this challenge, this paper proposes a novel neural network model based on the YOLOv7 architecture, named CS-YOLO. CS-YOLO incorporates several key technical innovations to enhance its detection performance. First, the model employs a dual-branch downsampling module (MSDown) to effectively address the issue of small crater feature attenuation during consecutive downsampling in the backbone network. Second, a multi-scale feature extraction module (MSFE) is designed to enhance the semantic information of small craters, significantly improving the model’s recognition ability in complex terrain. Finally, the cross-scale feature fusion module (CSFF) uses 3D multi-scale convolutions, not only extracting the structural features of small craters but also effectively integrating these features into the small head branch, greatly improving the model’s detection accuracy.Experiments were conducted using Mars solar thermal infrared remote sensing images, and the results show that CS-YOLO excels in detecting small craters, achieving a recall rate of 82.1% and an accuracy rate of 86.9%. Compared to other object detection algorithms, CS-YOLO demonstrates significant advantages in recognizing small craters in remote sensing images.

Phase unwrapping (PU) is a critical step in interferometric synthetic aperture radar (InSAR) data processing. Among all PU methods, multibaseline PU (MBPU) methods is a state-of-the-art method, which can overcome the limitations of the phase continuity assumption in the traditional single-baseline PU methods. However, the MBPU methods still cannot effectively balance PU accuracy and efficiency when processing large-size interferograms. In particular, the current best MBPU methods, two-stage programming approach (TSPA), cannot work well when the baseline ratio is less than 2, limiting its application. To solve this problem, a WaveCluster-based robust and fast MBPU algorithm (WCRFPU) is proposed in this paper. First, appropriate initial grid and neighborhood parameters are selected according to the exclusive information of the InSAR data set to reduce the number of wrong clusters caused by dimension mismatch. Then the WaveCluster algorithm is used to cluster the intercept map with 3D clustering features, which can obtain more accurate clustering results and efficiently handle large-size interferograms. Subsequently, a cluster correction step is added to improve the PU accuracy further. Theoretical analysis and experimental results show that this method has more advantages than the existing MBPU methods in efficiency, accuracy, and adaptability to baseline ratios when processing large-size interferograms.

Traditional synthetic aperture radar (SAR) image sea-land segmentation algorithms overlook the ship distribution priori-information provided by the automatic identification system (AIS) data, resulting in poor segmentation performance in complex environments such as ports, marine wetlands, beaches, and other sea-land boundaries. To address the above issues, this article comprehensively uses dual-polarization (VV and VH) SAR images and AIS data as the data source, and it specifically proposes a novel pyramid vision transformer (PVT) assisted by the long-time AIS data (AIS-PVT) for sea-land segmentation. AIS-PVT is the first attempt to integrate the ship distribution density priori-information, provided by the long-time AIS data, into the PVT network, thus the multiscale features of the sea and land can be better distinguished. In the decoding stage, we design a feature filter module (FFM). It aggregates features separately along two spatial directions from the skip connections, enhancing the representation of objects of interest while reducing the influence of redundant information. Furthermore, we develop a boundary-pixel-aware function to steer the model training process, allowing AIS-PVT to concentrate more on the neighborhood information of boundary pixels. Importantly, the AIS-PVT method captures global multiscale information and enhances the model’s data fusion capability. The conclusive experimental results demonstrate the superior performance of our approach in sea-land segmentation tasks, outperforming other state-of-the-art (SOTA) techniques.

Multi-baseline (MB) Phase Unwrapping (PU) technology is an important step in Synthetic Aperture Radar Interferometry (InSAR). The related method of MBPU aims to improve the noise robustness of the Chinese Remainder Theorem (CRT). As one of the popular methods of MBPU, the existing MBPU method based on cluster analysis (CA) has some challenges that need to be improved in practical applications. For example, the clustering results are inaccurate, the time required to process large-scale interferograms is too long, and the noise clusters in the clustering results are not processed further. To address the above issues, this paper proposes a WaveCluster-based fast and large scale MBPU method (WCFLS). Firstly, the InSAR data set is preprocessed, and then the WaveCluster algorithm is used to process the data set. Finally, the noise clusters in the clustering results are firstly identified accurately, and then the cluster numbers are reassigned to the noise clusters. Theoretical analysis and experiments show that the proposed method has a better improvement in noise robustness and efficiency.

Layover is a kind of geometric distortion in radar systems with side-look imaging, especially in mountainous and dense urban areas. It causes phase distortion and alters target characteristics in the acquired images, which directly hinders the application of radar images. In this letter, the multilayer feature fusion attention mechanism (MF2AM) is proposed to extract layover from interferometric synthetic aperture radar (InSAR) imagery automatically. First, the SAR image, the corresponding coherence map, and interferometric phases are channel-fused to enhance semantic information of layover areas. Then, the fused image is fed into MF2AM to extract the essential features of layover. Finally, the detection results are produced via MF2AM. MF2AM consists of the encoder and the decoder. The encoder contains three parts: the resnet101, attention-based atrous spatial pyramid (AASP), and the semantic embedding branch (SEB). In the decoder, step decoding is used to better fuse high- and low-level features and improve the effect of edge segmentation. To verify the proposed method, the images of millimeter wave InSAR system are used for the experiment, and the performance is compared with DeepLabV3+ and Geospatial Contextual Attention Mechanism (GCAM). The results show that the MF2AM has achieved obvious performance advantages. The average pixel accuracy and average intersection over union (IOU) are 0.9601 and 0.9310, respectively, and the average test time is only 7.97 s.

Although deep learning techniques have achieved noticeable success in aircraft detection, the scale heterogeneity, position difference, complex background interference, and speckle noise keep aircraft detection in large-scale synthetic aperture radar (SAR) images challenging. To solve these problems, we propose the geospatial transformer framework and implement it as a three-step target detection neural network, namely, the image decomposition, the multiscale geospatial contextual attention network (MGCAN), and result recomposition. First, the given large-scale SAR image is decomposed into slices via sliding windows according to the image characteristics of the aircraft. Second, slices are input into the MGCAN network for feature extraction, and the cluster distance nonmaximum suppression (CD-NMS) is utilized to determine the bounding boxes of aircraft. Finally, the detection results are produced via recomposition. Two innovative geospatial attention modules are proposed within MGCAN, namely, the efficient pyramid convolution attention fusion (EPCAF) module and the parallel residual spatial attention (PRSA) module, to extract multiscale features of the aircraft and suppress background noise. In the experiment, four large-scale SAR images with 1-m resolution from the Gaofen-3 system are tested, which are not included in the dataset. The results indicate that the detection performance of our geospatial transformer is better than Faster R-CNN, SSD, Efficientdet-D0, and YOLOV5s. The geospatial transformer integrates deep learning with SAR target characteristics to fully capture the multiscale contextual information and geospatial information of aircraft, effectively reduces complex background interference, and tackles the position difference of targets. It greatly improves the detection performance of aircraft and offers an effective approach to merge SAR domain knowledge with deep learning techniques.

Phase unwrapping (PU) and phase filtering are the key procedures for the interferometric synthetic aperture radar (InSAR) technology. As one of the most popular multibaseline PU (MBPU) algorithms, the cluster-analysis (CA)-based MBPU algorithm still has some problems that need to be improved. To begin with, the cluster ambiguity vector is obtained by searching the nearest integer point to the cluster centerline with known slope and intercept in the search space. It will be time-consuming and inconvenient when the number of baselines or the search space is too large. In addition, they do not have the capacity of phase filtering. Moreover, they do not consider the impact of different baseline combinations on the performance of the CA-based MBPU algorithm. For these reasons, a novel CA-based MBPU and filtering (MBPUF) algorithm is proposed in this article. The main contributions of this article are that it gives the closed-form solving formulas of the cluster ambiguity vector to improve the efficiency of the CA-based MBPU algorithm, proposes a novel MB InSAR phase-filtering strategy that makes the CA-based MBPU algorithm capable of solving the phase-discontinuity problem and improving the height-reconstruction accuracy simultaneously, and utilizes the optimal baseline combination to improve the robustness of the CA-based MBPU algorithm. Theoretical analysis and experiments on both simulated and real MB InSAR data sets show the effectiveness and robustness of the proposed closed-form robust CA-based MBPUF algorithm.

Long-term surface deformation monitoring of highways is crucial to prevent potential hazards and ensure sustainable transportation system safety. DInSAR technique shows its great advantages for ground movements monitoring compared with traditional geodetic survey methods. However, the unavoidable influences of the temporal and spatial decorrelation have brought restrictions for traditional DInSAR on the application for ribbon infrastructures deformation monitoring. In addition, PS and SBAS techniques are not suitable for the area where adequate natural high coherent points cannot be detected. Due to this, we designed an integrated highway deformation monitoring algorithm based on CRInSAR technique in this paper, the processing flow including Corner Reflectors (CR) identification, CR baseline network establishment, phase unwrapping, and time series highway deformation estimation. Both the simulated and real data experiments are conducted to assess and validate the algorithm. In the scenario using simulated data, 10 different noise levels are added to test the performance under different circumstances. The RMSE of linear deformation velocities for 10 different noise levels are obtained and analyzed, to investigate how the accuracy varies with noise. In the real data experiment, part of a highway in Henan, China is chosen as the test area. Six PALSAR images acquired from 22 December 2008 to 09 February 2010 were collected and 12 CR points were installed along the highway. The ultimate time series deformation estimated show that all the CR points are stable. CR04 is undergoing the most serious subsidence, with the maximum magnitude of 13.71[Formula: see text]mm over 14 months. Field leveling measurements are used to assess the external deformation accuracy, the final RMSE is estimated to be [Formula: see text][Formula: see text]mm, which indicates good accordance with the result of leveling.

In this paper, a high precision algorithm to estimate initial phase offset fast is presented for airborne dual-antenna InSAR system. Firstly, the factors influencing the initial phase offset are analyzed, and then the influence of the initial phase offset to DEM is given. According to the analysis, the algorithm of real-time estimation of the initial phase offset with high precision is presented. In order to evaluate the performance of the algorithm, two sets of real airborne dual-antennal InSAR data from the Institute of Electronics, Chinese Academy of Sciences (IECAS) are processed. The results prove the algorithm can get superior precision initial phase offset in real-time, which is very useful for real-time InSAR system.

The gradually increase of highway ground deformation has caused more disasters, which indicates high potential of traffic danger. In this paper, an integrated highway ground deformation monitoring algorithm based on CRInSAR is designed and carried out. Both the simulated experiment and real data experiment have been designed and implemented in order to validate the algorithm proposed. Due to the flexibility and good backscatter characteristic of CR in SAR images, the method shows high precision in deformation detection on highway object. The result of the simulated experiment shows the ideal feasibility of the algorithm, while that of the real data experiment shows good performance in highway deformation monitoring. The method proposed can play significant importance in the prevention of traffic accident induced by the accumulated highway ground deformation.

Phase difference measurement of sinusoidal signals can be used for phase calibration in spaceborne/airborne single-pass InSAR system. However, there are little discussions about the selection of the sampling frequency on the undersampling condition and corresponding measuring method when the signal's frequency is too high. In order to solve the problem aforementioned, this paper proposes a modified method based on Coherent Accumulation and DFT. Firstly, the appropriate under-sampling frequency is chosen to sample the two sinusoidal signals with the same frequency. Then, the sampled signals are coherent accumulated with the period of the baseband signal. Thirdly, the accumulated sampled signals are used to calculate the initial phases of the two sinusoidal signals by using the Discrete Fourier Transformation. Lastly, the phase difference is gotten from the initial phases. Experiment results prove the effectiveness of the method.