IEEE Geoscience and Remote Sensing Magazine - March 2020 - 31

a 3 × 3 patch. After formulating the joint pixel patch, SHP
identification and phase optimization are performed by the
time-series likelihood ratio test and the joint-subspace projection method, respectively. Using the coherence information from neighboring pixel stacks, JSInSAR is able to provide
reliable parameters estimation even in the presence of some
coregistration errors.
PERSISTENT SCATTERER PAIR
The most important feature of the PSP method is that it defines and uses PS pairs to identify PSs. It overcomes the misselection and omission caused by PS identification based
on amplitude and improves the accuracy of the parameter
solution [74], [75]. The core step of the PSP algorithm is to
establish the PSP network using an iteration method. In the
iteration, the deformation rate and elevation correction are
estimated by maximizing the phase coherence. When the
coherence coefficient is greater than the given threshold,

ux(5)

the PS pair is added to a set. This process is repeated until
no new point pairs can be added. At the same time that the
iteration is completed, the parameters of the entire scene
are obtained. Table 1 compares of various time-series InSAR techniques.
OTHER MATTERS
In addition to the complete time-series InSAR frameworks mentioned previously, there are also some minor
improvements to the time-series InSAR technique. In
[76], the periodic deformation caused by seasonal temperature variation can be solved on the basis of the PSI
technique. Some methods consider high-order deformation models, such as deformation acceleration [77].
In [78]-[80], a temperature model incorporated into the
phase model to estimate the thermal expansion component of the building is considered. To avoid local extremums, the authors in [81] applied a particle swarm

SLC 1

SLC k

SLC N

1-1 2-1 3-1

1-k 2-k 3-k

1-N 2-N 3-N

4-1 5-1 6-1

4-k 5-k 6-k

4-N 5-N 6-N

7-1 8-1 9-1

7-k 8-k 9-k

7-N 8-N 9-N

1-1

1-k

1-N

5-1

5-k

5-N

9-1

9-k

9-N

FIGURE 13. The formulation of a joint pixel vector using a 3 × 3 patch. SLC: single look complex.

TABLE 1. A COMPARISON OF TIME-SERIES INSAR TECHNIQUES.
METHOD

INTERFEROGRAM
MODE

MEASUREMENT
POINT

DEFORMATION MODEL

PHASE UNWRAPPING

OPTIMIZER

PSI

Single reference

Linear model in time

1D time

Periodogram

SBAS

Multireference

Linear model in time

2D space

MCF + SVD

CPT

Multireference

PS and DS

Linear model in time

1D time

Conjugate gradient

PSP

Single reference

Linear model in time

1D time

Periodogram

SqueeSAR

Single reference

Linear model in time

1D time

Periodogram

JSInSAR

Single reference

Linear model in time

1D time

Periodogram

StaMPS

Single reference

Spatial smoothness

2D space + 1D time

Filtering

QPS

Multireference

PS and DS

Linear model in time

1D time

Least square

TCP-InSAR

Multireference

Linear model in time

None

Least square

MTDA

Multireference

Spatial smoothness

2D space + 1D time

MCF

STUN

Single reference

Linear model in time

2D space + 1D time

LAMBDA

LAMBDA: least squares ambiguity decorrelation adjustment.

MARCH 2020

IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE

IEEE Geoscience and Remote Sensing Magazine - March 2020

Table of Contents for the Digital Edition of IEEE Geoscience and Remote Sensing Magazine - March 2020