In most of the current operational autonomous underwater vehicles (AUVs), a survey mission is generally composed of two main stages. The first one conducts an exhaustive coverage over an area of interest, while gathering data of the sea bottom. Then, and after processing the collected data, a second mission is programmed to obtain more detailed information of potential targets, and to cover the gaps that resulted from the first exploration. However, this two-survey strategy can be inefficient, since it requires establishing a communication link between the AUV and its operator for retrieving the data and reprogramming the second mission. To cope with this situation, we present a mission planner that endows an AUV with the capability of extending its missions online. With our approach, the vehicle is also required to conduct an initial and predefined survey of an area of interest, but it processes the gathered data onboard to plan 3D feasible paths to complement the initial exploration. To validate our approach, we present real-world results with the AsterX AUV.