Nanoproducts represent a potential growing sector and nanobrous materials are widely requested in industrial, medical and environmental applications. Unfortunately, the production processes at the nanoscale are difficult to control, and produced artifacts often exhibit local defects that prevent their functional properties. We present a fully-automated solution to detect defects in nanobrous materials during their production, yielding smart-manufacturing systems that reduce quality-inspection times and wastes. We analyze SEM images of nanobrous materials and learn, during an initial training phase, a model yielding sparse representations of the structures that characterize correctly produced nanobers. Defects are then detected by analyzing patches in test images and assessing the goodness-of-t of each patch to the learned model. The proposed solution has been successfully validated over 45 images acquired from samples produced by a prototype electrospinning machine. The low computational times indicate that the proposed solution can be adopted for real-time monitoring in an industrial-production scenario.