We present a simulation study on the peeling of carbon nanotubes bundles interacting with a flat substrate, represented by the back surface of an atomic force microscopy cantilever. A defected sample, acquired in situ using a scanning electron microscope, was investigated under different peeling configurations by finite element method simulations. The coupled computational-experimental analysis let to identify the position and the entity of a structural defect by means of reverse image correlation problem. By exploiting this defective fibre it was also possible to quantify, as indirect measure, the friction and adhesion forces between the bundle and the substrate of few pN magnitude, otherwise difficult to measure with the resolution of currently available instruments. The proposed approach can be useful to study the tribology-induced mechanical behaviour of one-dimensional nanostructures as well as for real-time identification and monitoring of nanodefects for industrial applications, such as nanoelectronics.