We present an error tolerant path planning algorithm for Micro Aerial Vehicle (MAV) swarms. It is GPS-free. The MAVs find their way using cameras to identify a series of visual landmarks. The landmarks lead towards the destination. MAVs are unaware of the terrain and landmark locations. Landmarks hold a-priori information whose interpretation is prone to errors. We distinguish two types of errors: recognition and advice. Recognition errors are due to misinterpretation of sensed data or a-priori information, or confusion of objects. Advice errors are due to outdated or wrong information associated to the landmarks. The MAVs cooperate and exchange information wirelessly, to minimize the errors. Consequently, the swarm experiences data quality amplification and error reduction. Quality amplification is related to the number of MAVs. The solution effectively achieves an adaptive error tolerant navigation system.