In response to the comments of the form "just install PTZ cameras", I have done an installation like this for a modest-sized site, and it is extremely difficult and time consuming (+ expensive) to install enough cameras to cover the site if you want any chance of actually recognising a face or a numberplate. If all you want to know is: "blob detected at rear gate", it's of course a lot easier, but then you might as well use a PIR sensor.
The problem is mapping a 3D space, where a 'perp' can often face in any direction, onto a series of 2D planes. PTZ cameras only help if they are meticulously programmed to track movement or respond to motion events from other sources, or you have a skilled operator 24x7.
It's even harder if the system needs to work well at night, since it requires either IR or visible light illumination.
I don't know how well this drone+sensors setup would work, but can definitely imagine there is a subset of the market where this would be better than fixed cameras, perhaps a large property where there are multiple entrances and exits, and lots of potential targets. e.g. a farm, where there are long easily traversed boundaries between the property and neighbouring properties.
The problem is mapping a 3D space, where a 'perp' can often face in any direction, onto a series of 2D planes. PTZ cameras only help if they are meticulously programmed to track movement or respond to motion events from other sources, or you have a skilled operator 24x7.
It's even harder if the system needs to work well at night, since it requires either IR or visible light illumination.
I don't know how well this drone+sensors setup would work, but can definitely imagine there is a subset of the market where this would be better than fixed cameras, perhaps a large property where there are multiple entrances and exits, and lots of potential targets. e.g. a farm, where there are long easily traversed boundaries between the property and neighbouring properties.