Structural cracks reduce the performance and service life of buildings, bridges, and aircraft structures, leading to catastrophic failures resulting in economic losses and fatalities. To avoid such consequences, regular health monitoring and maintenance is required, especially for critical structures that carry high levels of dynamic and fatigue loading and whose failure would be catastrophic. Many techniques are available for structural health monitoring, including visual inspection, ultrasonic testing, acoustic emission, magnetic field, and vibrational-based methods. Any damage causes degradation in the natural frequencies and vibration modes of a structure, which are considered in vibration-based methods to characterize the damage. The focus of this research is to develop a more efficient method for the detection and characterisation of arbitrarily oriented surface cracks in isotropic plates, in terms of five parameters, namely the longitudinal and transverse location, length, depth and orientation. To achieve this objective, an analytical solution based on strain energy is used to generate synthetic data that quantifies changes in the natural frequencies for different crack locations and intensities based on noise-free simulation. The inverse problem, i.e. the determination of the crack parameters based on measured changes in natural frequencies can then be solved based on the use of synthetic data with a gradient based optimisation technique.