Radar interferometry (InSAR) is a valuable tool to measure surface motion. Applying time series techniques such as Persistent Scatterer Interferometry (PSI), InSAR is able to provide surface displacements maps with mm-precision. However, InSAR can still be further optimized, e.g. by exploiting spatial characteristics of the signal of interest. This study addresses surface deformation associated with geological faults. In principle, this signal is generally spatially smooth but with significant-to-large gradients at fault locations. We first focus on optimizing InSAR time series analysis, in particular the PSI method, for this specific class of ground deformation processes. Secondly, we apply the improved technique to study fault-related motion in the southern Netherlands, with special interest in detecting neotectonic motion in the Roer Valley Graben and deformation in the abandoned mines of South Limburg. The proposed optimization adapts PSI to analyze in an iterative manner the signal of interest to estimate spatially the probability density function of displacements. Since the signal is expected to change quickly near faults, we do not restrict this distribution to be unimodal but we allow it to have any shape. Finally, we use the determined distribution to constrain, through Bayesian inference, phase unwrapping (the operation of unfolding the phase outside its natural range of (?pi, pi] radians). We demonstrate a substantial benefit of the Bayesian approach as we show a decrease in the number of unwrapping errors. This thesis also suggests a method that analyzes interferometric phases to estimate noise variance. It is built upon the assumption that coherence can be spatially correlated. The estimated stochastic parameters are used in phase unwrapping by assigning lower weights to noisy observations. The improved methodology is applied to study fault-related motion in the southern Netherlands, exploiting data from three satellite missions: ERS-1, ERS-2 and Envisat. In particular, we focus on two main areas: the Roer Valley Graben and the abandoned mines of south Limburg. In the Roer Valley Graben area, a deformation signal associated with geological faults is detected. However, we do not observe any significant indication to atribute a tectonic origin to this signal for two main reasons. First, during large part of the studied period the most of the graben uplifts with respect to adjacent horsts at rates of ~1 mm/yr, behaving opposite to predicted by tectonics. Second, the deformation signal in this area appears to be largely related to water pumping. For example, we observe an uplift signal of about +4 mm/yr that matches in time and space with the cease of pumping in the Erkelenz Coal District, which is located in the Peel horst, adjacent to the Roer Valley Graben. Concerning the mines of South Limburg, we detect strong surface displacements (uplift) which appear to be centered on the old mines and constrained by tectonic faults. The signal is variable in space and time, with uplift rates up to 20 cm in 18 years, and relatively large gradients across faults (~5 cm/km), in the same time span. Laterally the uplift signal propagates towards the west in this period. The comparison of surface displacements with rising groundwater levels reveal a strong correlation between the two, suggesting the groundwater to be the cause of the uplift. Assuming that rising ground water levels in the abandoned mines are responsible for the uplift, we estimate the relation between the groundwater and the associated uplift. The skeletal storage coefficient, which directly depends on porosity, is on average 0.5±0.1·10?3, implying that 1 m of water level increase produces 0.5 mm uplift. As we expect that the water may rise many tens of meters, especially in the western side, this may result in several additional centimeters of future uplift. Essentially, the surface displacements that we observe in the southern Netherlands seem to be mainly caused by fluctuations in groundwater flow, which appear to be constrained by faults.