数据和所有PS点都是有效的。在PS处理中的精度因子的计算方法和GPS系统的精度因子一样，考虑到时序图像的几何，基线估算和相应的2π高程模糊，还会用到默认的PS点密度。
最后，当PS处理结束后，会计算出每个PS点的实际精度，基于象元相干性和实际的PS密度
Q. -如何知道哪些数据是因为配准失败而被丢弃的？
A. –当配准失败的时候，在输出路径下会生成相应的文件： \，文件中会将由于配准问题而丢弃的数据列出来。
Q. –在PS分析完成之后，有没有哪些软件可以检查的关键要素，确保结果的可靠性？ A. – 其中一项就是检查相关的配准过程：\文件包含了那些没有配准的图像。另一项要检查的是大气影响的去除，根据经验来说，对比\（大气影像去除之前）和\（大气影响去除之后），多时间相干性增加，说明大气校正是成功的。
> 对比前后的cc文件，可以更好的确定如何设置滤波器参数（低通，以米为单位；高通，以天为单位），以避免相干性的损失，来更好的估算形变量。
Q. - SBAS 的首字母分别地表什么？ A. – 代表 “small-baseline subset”，该技术最早起源于：Berardino, Fornaro, Lanari and Sansosti (参看参看文献).
Q. – 在SBAS连接图生成的时候，最大最小的空间和时间基线的确定有什么标准？
A. – 设置最小和最大基线一般是根据数据处理的目的来确定的，看是要用SBAS做形变监测还是提取DEM。如果是要做形变监测，典型的一个设置就是能得到所有的连接图，反之，如果要生成DEM，同样是不连贯的像对可以得到高精度的DEM。
基于上面的叙述，如果在生成DEM的时候选择了“允许不连贯的区域”（Allow Disconnected Blocks），最大最小基线就可以设置为可去除最小像对基线的值，比如临界基线的20%的像对就去除了，这些是对高程测量无用的数据。
另一方面，不管是提取DEM还是测量形变，最好避免选择太大基线的像对，这样的数据相干性很低，这种情况下，空间基线的上限可以设置为临界基线的50%，时间基线的上限也可以参考这样的设置，因为相干性和基线距离是成反比的，影响时间失相干的因素和地表覆盖有关的，如植被、土壤含水量、积雪覆盖等。
在处理中会考虑到这些阈值，在处理过程中是公开的，在干涉生成这一步之后，分析结果以确定在连接表编辑的时候去除哪些像对。
Q. – 在SBAS反算的时候，如何设置大气的低通（Atmosphere Low Pass）和高通（High Pass）滤波参数？
A. – 窗口大小越小，滤波器就越强，滤波器中引进的平滑可以评估，对比\中的时间信号（没有进行大气去除的结果）和最终的形变结果。这有助于理解是否大气滤波平滑去
除了重要的形变信息。这可以发生在那些大气模式（atmospheric patterns）很少的情况下（如山区范围）因此滤波大小必须要相应的调整，这可以通过在SBAS反算之前观察干涉图评估出来。然而，如果形变模式（displacement patterns）窗口就很小的话（等于或小于大气的），这种情况下不能过分的减少窗口，以防止将形变作为大气信息去除掉了。
Q. –软件如何处理数据集的低相干性的散射区域？
A. - The example below shows a typical case of scattered coherence areas. The area of interest, where subsidence phenomena have been reported, is in the lower right corner.
This area is well represented and the coherence is good, but the surrounding low coherence zones are can affect the SBAS inversion process and eventually the measurement accuracy. However, if there are enough interferograms (theoretically at least five per acquisition) the program is able to \the missing information by means of the 3-dimensions unwrapping approach.
Q. – 干涉图系列的解缠方法选择的是什么？
A. - Actually the unwrapping execution depends on the Interferometric Stacking approach which is adopted:
◆ in the Persistent Scatterers, based on the original publication of Ferretti et al., the unwrapping is not performed for the estimation of the displacement rate and height corrections, since a pixel-wise spectral analysis approach in the time-baseline plane is exploited; this approach has the advantage of avoiding the need of unwrapping by working on the complex data only.
◆ in the SBAS Inversion, it is possible to choose between two main methods: Region Growing or Minimum Cost Flow, this last one either with a square or with an irregular triangulated - Delaunay - grid (Default Values>Interferometry). In case the Delaunay method is adopted, a 2- or 3-dimensions unwrapping (the third dimension being represented by the time) can be selected. At this regards it must be noted that the 3-dimensions approach provides superior results when there are disconnected areas (typically due to low coherence), as it exploits the high coherence
interferograms (third dimension) to estimate how to create new connections in \coherence) interferograms; the disadvantage of this method is that it is much more costly in terms of processing time. The unwrapping, in the SBAS processing chain, is carried out two times: once before a first estimate of average displacement rate and height correction and once more afterwards to refine the first results.
Q. – 设置分解级别的时候，如何设置最优的级别，有无特定的依据？ A. –这里没有特定的规则，因为该参数的最优设置取决于区域的相干性和解缠的方法。 There are not specific rules since the optimal setting of this parameter depends strongly on the scene coherence and also on the unwrapping method adopted. 大多数情况下，尤其是选择了默认的方法，分解等级为1是最好的设置。有时候，将等级设置为1（1A – 干涉生成工作流这一步）增加到2（在第一次反算这一步），等级的增加会得到更好的结果。
Q. - After the SBAS inversion, some of the Unwrapped Interferograms are still affected by Directional Slopes. Is this due to a problem during the Refinement and Re-Flattening step? What is the best way to address such issues?
A. - The first thing to do it is to visualize and verify the unwrapped data (_upha_list_meta), after the Interferometric Workflow step, in order to understand where the GCPs must be located for the next Refinement and Re-Flattening. The worst pairs (i.e. very low or scattered coherence, which typically causes a bad unwrapping) shall be removed by means of the appropriate tool; if possible before the Refinement and Re-Flattening, otherwise after the First Inversion step.
After the execution of the Refinement and Re-Flattening, in order to ensure that the GCPs have been properly selected (in terms of position, distribution and quantity), another visual analysis is needed to verify that major residual phase ramps are not in; if they are still present, a higher number of GCPs (20 or more) is probably required.
Once the re-flattened data have been checked, the First Inversion step can performed, which generates a new set of the unwrapped data (these are stored in the \After this step, the remaining \inversion.
Note that the program is implemented in a way that, even when some small residual ramps remain (of course the less the better...), the inversion process is not notably affected.
Q. – 有没有生成轨道GCP文件的简单方法？
A. – GCP点当需要输入\时，该参数是最重要的参数，其选择标准非常重要。在干涉测量模块的QA里面已经提到过了，而在干涉叠加模块这里，需要增加说明的是： it is that the same set of points must be used to \have been created after the coregistration onto the \not impossible...) to find the best point configuration for all pairs at once, the suggested
approach it is to try a point distribution which is good for the majority of the pairs (for this purpose it becomes useful to visualize the stack of images by means of the meta file generated automatically by the program); in several cases the use of the Default Values>Flattening>Refinement and Re-flattening>Refinement Method>Residual Phase, which allows the use of the less precise but most robust phase correction approach, shall be adopted. It must be also taken into account that some points, which are inside the imaged area for some pairs, can fall outside in others and thus it can be required to enter more points; in other cases, for instance when the original orbits are all very much accurate and there is not any visual evidence of residual phase ramps, it can be sufficient to simply remove a phase offset (constant value) by choosing a Default Values>Flattening>Refinement and Re-flattening>Refinement Method>Residual Phase Poly Degree equal to 1.
In any case, also relying on very precise orbits, the use of the \the reflattening process must be always executed in order to correct the phase offset (i.e. constant phase removal).
Q. – 对SBAS结果最好的分析方法是什么？
A. – 得到的结果是以meta文件分组存放的 ，为了方便做多时相的时间序列分析。Meta文件可以用任意栅格分析的工具，或者转换成矢量文件后，用矢量数据的分析工具进行分析。