This talk presents a panorama of my research related to the two-dimensional linear elasticity system. The first part is concerned with a geometric inverse problem: the identification of voids under Navier's boundary conditions (i.e. the elastic solid can slide in tangential direction while in the normal direction the displacement is clamped) from the knowledge of partially over-determined boundary data. Sensitivity analysis methods (shape derivative, topological derivative) are developed to spot numerically the flaws. Secondly, a parametric inverse problem is studied: the reconstruction of interface stiffness parameter (i.e. the interface tractions are continuous while the displacement is discontinuous across the debonded region and proportional to the interface traction). Lipschitz stability estimate is established and based on a new Carleman's inequality with suitable weight functions. Finally, I am interested in quantifying the elastic properties of intensely fractured rocks around tectonic faults. The density and complexity of the natural fracture networks over a wide range of spatial scales is modeled by a statistical scaling model calibrated with field observations and measurements. The effective parameters of the medium are estimated by the stochastic homogenization method.