Preventive conservation is a vital practice in Cultural Heritage that consists in the constant monitoring of the state of conservation of an artwork, with the final goal of minimizing the risk of damages and thus, to reduce the need of restorations. In this work, we propose a probabilistic approach for analysing alterations appearing on varnished wooden surfaces (such as those of historical violins) based on an a-contrario framework. Our method works on time series of images, it is robust to noise and avoids parameter tuning as well as any assumption about the quantity and the shape of the worn-out areas. Furthermore, the proposed algorithm is adapted to the context of preventive conservation, with few data samples and imprecise annotations. As test set, we used image sequences included in the ``Violins UVIFL imagery'' dataset. Results illustrate the capability of the proposed method to distinguish altered areas from the surrounding noise and artifacts.