Underwater image formation is degraded by several factors, which causes the ocean to be a challenging environment for image processing. This paper aims to improve the visual servoing capability of an autonomous underwater vehicle by using pre-processing algorithms to improve the image quality. We used artificial fiducial markers to feed the visual controller. Therefore, three different methods for imaging enhancing were applied to the raw image aiming to increase the detection rate of the marker detector. Since the performance of the visual controller also depends on the detection time, this was also considered in the comparison. Finally, the algorithm that caused the best improvement in marker detection was tested in a visual servoing mission. The proposed methods have shown a significantly improvement in the marker detection rates and reasonable detection times for visual servoing of autonomous underwater vehicles. In terms of visual servoing missions, this work shows that the proposed methods not only increased the controller frequency, but also made visual servoing possible when water condition is not favorable and no marker can be detected without a pre-processing layer.