There are several options for augmenting the descent and landing phases in order increase the safety margin, including:
The 'man-in-the-loop' option of introducing manual/computer- assisted ground support and intervention into the navigation loop, based on visual information from onboard wide-angle camera(s), is particularly interesting as it would allow:
Timewise, ground intervention during the navigation process certainly seems possible, given that closure of the navigation loop (based on visual landmark recognition) would take only 5-10 seconds.
All image processing would be carried out on the ground, providing an excellent opportunity for the public to share in the high drama of the actual landing. A hovering phase, lasting about 2 minutes, might also be envisaged during the descent from 100 m to touchdown, to further facilitate obstacle avoidance.
The possibility of including a 'hopping' capability is also being considered as a fuel-efficient means of changing the landing spot should a hazard be found at the original target site.