While it's an interesting idea, it's not just possible with today's tech.
Black rhinos are only one subspecies, white rhinos are just as threatened by poaching and there are around 20 000 of those, so you're looking at around 25 000 rhinos in Africa.
You'd need at least three UAVs per rhino, one on station at any given time, one ready to replace it once its endurance runs out and one spare in case something happens to one of the other two. Then there's the problem of actually tracking the rhino 24x7, through night and bad weather, while ensuring that the base station and recovery team stays close.
You'd also need skilled pilots / operators to keep the UAV on station and interpret what's on the screen and, once again, you'd need more than one per rhino to enable 24x7 shifts.
Once you factor in the number of UAVs, the dispersed recovery and handling teams and the need for skilled operators, you could potentially be speaking of 75 000 UAVs and a similar number of teams and pilots / operators. Even if we assume that UAVs with sufficient endurance for this task could be reduced to $500 000 each, that still results in a capital cost of nearly $40 billion, over and above the sure-to-be-mammoth operating costs. So it's not feasible.
A better approach is to fit each rhino with a GPS tracker, which is being done, and to use ranger patrols augmented by technology like UAVs and ground radar to attempt to spot poachers moving in the parks. That's why this is where SANParks and others are focusing their attention and resources.