It depends on what exactly you want to do. Only optical mapping with gps referenced pictures? Or 3d mapping with lidar, radar or other sophisticated sensors? Multicopter are a bit easier to position but they are strictly payload and flight time restricted. Fixed wing have a lot more endurance but can't be postioned that exactly epecially for a longer time as they can't hover. How big is the area you want to map?
I agree with Bastians response, I would further add that it is essential to first fully define what you expect out of your project i.e. what product, map or DEM, what is its final use and map scale, accuracy, is there need to validate the map or DEM accuracy on completion, how important is a feature location on your map in terms of its xyz coordinates compared to ground truth? The more accurate the map or DEM you want to create then necessitates consideration of other factors e.g. how do you account for effects of tilt, yaw, roll of your mapping vehicle. What accuracy GPS is used, are ground stations to be used to improve GPS solution? Correct photograph GPS coordinates without accounting for vehicle movement will complicate consistency of final solution. Simple example, your mapping vehicle taking photogaphs at predetermined positions experiences an updraft which changes its elevation. The effect of this is a change in scale and area of photograph. Does your method of analyzing your data to obtain your final product account for the many variables that occur as a result of the process you decide on. Basic photogrammetry requires knowledge of ground xyz points identifiable on the photographs for bundle adjustment, is this necessary for your project to reduce error of the final product? These are some considerations necessary when planning on producing a map or DEM? There may be more, I here wanted you to have a slightly broader picture or get you to see from a higher perspective what you are dealing with. Hope this helps.