Fusion of GPS and InSAR data to derive robust 3D deformation maps based on MRF L1-regularization

Ji, Panfeng; Lv, Xiaolei; Dou, Fangjia; Ye, Yun

doi:10.1080/2150704x.2019.1701722

Cited by 5 publications

(6 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…12(b), (d) and (f). The errors of InSAR are calculated by equation (2). The general deformation trend is consistent with that obtained by IAUE-LSC ( Fig.…”

Section: B Real Casesupporting

confidence: 82%

“…On the one hand, InSAR can obtain spatially continuous deformation, but it is only one-dimensional (in the direction of the line of sight, LOS). On the other hand, GNSS can acquire high-precision three-dimensional displacements, but it is spatially discrete [2]. Obviously, they are highly complementary and can be combined to obtain highprecision three-dimensional deformation maps.…”

mentioning

confidence: 99%

“…where D(i, j) is the displacement observation of the corresponding observation point in the window,D(i, j) is the average value of the deformation observation of all points in the window, N is the number of pixels in the window. A priori weighting method has been used many times to obtain the optimal three-dimensional results, such as the analytical optimization method [3], the BFGS method [8], the ant colony optimization method [9], the MRF-L1 regularization [2], etc. A priori method may be reasonable in a sense, but it is not accurate.…”

mentioning

confidence: 99%

See 2 more Smart Citations

Applying InSAR and GNSS Data to Obtain 3-D Surface Deformations Based on Iterated Almost Unbiased Estimation and Laplacian Smoothness Constraint

Chen

et al. 2021

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

Self Cite

View full text Add to dashboard Cite

GNSS and InSAR data are integrated to extract the 3-D surface deformations, which are of great significance for studying geological hazards. In this study, two major problems are focused on integration. For one thing, we propose an Iterated Almost Unbiased Estimation (IAUE) method to estimate the variance components of GNSS and InSAR for the case where the estimation of variance components of multi-source data by traditional variance component estimation (VCE) methods may be negative and inaccurate. For another, considering that heterogeneous data errors may lead to unstable three-dimensional solutions, we propose adding the Laplacian smoothness constraint (LSC) to the function model, which can smooth the solutions by minimizing the second derivative of the displacements. These two methods are abbreviated as IAUE-LSC. In the simulation experiment, the performance of traditional Helmert variance component estimation (HVCE) is first compared with IAUE. IAUE can not only converge more quickly but also avoid negative variances. Furthermore, we find that the excessively large relative error ratio between GNSS and InSAR is an essential factor leading to the instability of the three-dimensional solutions. IAUE-LSC method is immune to this instability and can obtain more stable results. In addition, the 2018 Hawaii case demonstrates that IAUE achieves improvements of 2.58 cm, 2.77 cm, and 7.69 cm in the east, north, and up directions relative to the traditional weighted least squares (WLS) method. While the combined IAUE-LSC achieve improvements of 2.29 cm, 0.32 cm, and 1.68 cm compared to the IAUE alone.

show abstract

“…12(b), (d) and (f). The errors of InSAR are calculated by equation (2). The general deformation trend is consistent with that obtained by IAUE-LSC ( Fig.…”

Section: B Real Casesupporting

confidence: 82%

mentioning

confidence: 99%

mentioning

confidence: 99%

See 1 more Smart Citation

Applying InSAR and GNSS Data to Obtain 3-D Surface Deformations Based on Iterated Almost Unbiased Estimation and Laplacian Smoothness Constraint

Chen

et al. 2021

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

Self Cite

View full text Add to dashboard Cite

show abstract

“…Traditionally, the method of unifying the results of InSAR and GNSS data is to correct the GNSS displacements to the relative reference frame as InSAR. When this method is adopted, a certain GNSS point with small displacements that is far away from the large deformation area, is selected as a common reference point for InSAR and GNSS [12]. This method is generally applied in the combination of DInSAR and GNSS to monitor large deformation scenarios, such as earthquakes, volcanoes and landslides.…”

Section: Correction Of Insar Observationsmentioning

confidence: 99%

“…The research results largely fall into three categories. The first type of method is represented by the analytical optimization method [8,9], the BFGS method [10], the ant colony optimization method [11] and the MRF-L1 regularization method [12].…”

Section: Introductionmentioning

confidence: 99%

Deriving 3-D Surface Deformation Time Series with Strain Model and Kalman Filter from GNSS and InSAR Data

Wang

2022

Remote Sensing

Self Cite

View full text Add to dashboard Cite

This study proposes a new set of processing procedures based on the strain model and the Kalman filter (SM-Kalman) to obtain high-precision three-dimensional surface deformation time series from interferometric synthetic aperture radar (InSAR) and global navigation satellite system (GNSS) data. Implementing the Kalman filter requires the establishment of state and observation equations. In the time domain, the state equation is generated by fitting the pre-existing deformation time series based on a deformation model containing linear and seasonal terms. In the space domain, the observation equation is established with the assistance of the strain model to realize the spatial combination of InSAR and GNSS observation data at each moment. Benefiting from the application of the Kalman filter, InSAR and GNSS data at different moments can be synchronized. The time and measurement update steps are performed dynamically to generate a 3-D deformation time series with high precision and a high resolution in the temporal and spatial domains. Sentinel-1 SAR and GNSS datasets in the Los Angeles area are used to verify the effectiveness of the proposed method. The datasets include twenty-seven ascending track SAR images, thirty-four descending track SAR images and the daily time series of forty-eight GNSS stations from January 2016 to November 2018. The experimental result demonstrates that the proposed SM-Kalman method can produce high-precision deformation results at the millimeter level and provide two types of 3-D deformation time series with the same temporal resolution as InSAR or GNSS observations according to the needs of users. The new method achieves a high degree of temporal and spatial fusion of GNSS and InSAR data.

show abstract

Fusion of GNSS and InSAR time series using the improved STRE model: applications to the San Francisco Bay Area and Southern California

Yan

Dai²,

Xie³

et al. 2022

J Geod

View full text Add to dashboard Cite

Fusion of GPS and InSAR data to derive robust 3D deformation maps based on MRF L1-regularization

Cited by 5 publications

References 22 publications

Applying InSAR and GNSS Data to Obtain 3-D Surface Deformations Based on Iterated Almost Unbiased Estimation and Laplacian Smoothness Constraint

Applying InSAR and GNSS Data to Obtain 3-D Surface Deformations Based on Iterated Almost Unbiased Estimation and Laplacian Smoothness Constraint

Deriving 3-D Surface Deformation Time Series with Strain Model and Kalman Filter from GNSS and InSAR Data

Fusion of GNSS and InSAR time series using the improved STRE model: applications to the San Francisco Bay Area and Southern California

Contact Info

Product

Resources

About