Identification of stable pointwise target is an important procedure in multi temporal multi-temporal INSAR analysis and application in monitoring regional surface deformation. The accurate identification result helps to improve the land subsidence inversion precision.Various methods for pointwise target detection have been proposed during the past two decades from different respective. The methods can be divided into some main categories according to the criteria for coherence point selection, such as amplitude dispersion index DA, signalto-Clutter ratio and phase stability. The DA method performs a time series analysis on amplitude instead of phase, and reflects the stability of series amplitude. The advantage of coherence point selection by phase stability can identify some special objects with a stable phase, which further increases the density of the stable pointwise target points, but ignores the highly scattering reflection characteristics of the coherent point. The existing methods take insufficient account of the overall features of stable pointwise targets. For ensuring stable scattering mechanism and temporal stability of pointwise target, an improved method with subaperture correlation was proposed in this paper. First, the subaperture correlation properties IHP of SAR images were obtained by spectral decomposition. Then the stability of targets is evaluated based on series two-aperture spectral coherence, by which the coherence points with high scattering could be identified and detect as PSC1.The DA threshold is utilized as the second criterion, which means any pixel in PSC1 with amplitude dispersion less than 0.4 can be determined as PSC2 Then phase stability analysis was carried out to screen out the true stable points from PSC2 with the Characteristics of high scattering mechanism and temporal stability. The experiments of stable pointwise target detection were performed using 33 high resolution SAR images collected by the Terra SAR X-band radar sensor covering Beijing. The detection results demonstrated that the improved method can detect more accurate and reliable pointwise targets than those identified by traditional methods To further confirm the effectiveness of the proposed method, the small subset INSAR technique based on the proposed coherence point detection method was adopted to retrieve the ground deformation by 40 scenes dataset acquired from 2003 to 2009 in Beijing. The vertical surface displacement rates during this period was validated by the leveling observations, with RMSE =1.36 mm/a, indicating two types of subsidence matched very well. The maximum subsidence rate of Beijing in investigated area has reached –92.25 mm/a, with an obvious uneven spatial distribution. Subaperture correlation is sensitivity to the high scattering body and can ensure stable scattering mechanism and temporal stability of pointwise target. Both coherence point detection results and the primary surface deformation proved the effectiveness of the proposed method. The deformation result during 2003–2009 has undergone severe land subsidence with high spatial aggregation characteristic, and the regional subsidence and the groundwater exploitation reveal good corresponding relationship, the more exploitation, the higher deformation rate.

Subaperture decompositions;high coherence point detection;small subset INSAR;Temporal-spatial distribution;ground water