The capture of vortical structures in high-fidelity CFD simulations for complex geometries is of great interest in the modelling of vortex-induced vibration problems, and for industrial aerodynamic design, such as F1 car or aircraft aerodynamics. This works aims at developing an adaptive methodology to capture vortical structures. Its main components are a feature-based error indicator that will identify the regions occupied by vortices, where additional mesh resolution is required, and a mesh modification strategy to locally increase resolution. We will first present a review of vortex identification methods and their performance appraisal using the simulation of the von Kármán vortex street behind a cylinder at a sub-critical Reynolds number as the benchmark. The set of best vortex indicators identified will then be used to increase the resolution of high-order hp/spectral simulations of incompressible flows via available mesh adaptation modalities. Specifically, we will combine h-adaptation by boundary-conforming iso-parametric split, and r-adaptation using a variational framework to redistribute the available degrees-of-freedom. These methods are implemented in the open-source high-order mesh generation tool NekMesh. The methodology will finally be applied to capturing vortical structures in more complex geometries such as a NACA wing tip and a F1 car front wing.