
DJI just came out with the phantom 4 their latest camera platform quad
In common parlance anything like this is called a drone but that typically implies capabilities these things don't have.
So the racing quads both Rob and myself own would be called a drone by the man in the street but they have almost no autonomous features (self level and that's about it and most of the time that's turned off). By comparison something like the phantom has self level and an array of extra positional sensors including gps as well as the ability to fly autonomously without input from the pilot.
In the case of this new phantom calling it a drone is probably accurate. It has the gps autopilot features of the previous models and the ultrasonic ground slip sensors that allow it to maintain a steady position even in fairly high winds. The big addition is on board collision avoidance, it has 4 additional cameras, as well as the gambled 4k main camera, that it uses for obstacle avoidance. It can identify collision objects and if commanded to fly at them it can avoid or just refuse to go near them.
It can also use the cameras to identify people/other objects and then track them in 3d space. This is part of a new feature called tap fly where the conventional dual stick controllers are replaced by a simple tap interface you tap to identify what the target of the scene is and then tap to give it a direction to fly and it then handles the rest.
The technology in these machines is developing in leaps and bounds although the legal situation surrounding them is getting more fraught. The US have a registration scheme already and after recent incidents at heathrow they are pushing for the same in the uk. It's an interesting case of technology advancing so fast that the laws and society are struggling to keep up with it.
An impressive new machine from DJI who are really the industry leader in this field of autonomous aerial photography platforms
Comments
I completely agree with the above.
Our quads have more in common with radio controlled planes than drones. I saw a joke on a FB group I'm part of where someone said "Not even the trees can stop a fly-away now", which made me laugh.
DJI are streets ahead of the competition. Parrot is probably its nearest neighbour (they are both stocked in the Reading Apple store) but this sense-and-avoid stuff is mightily clever.
This is really getting towards the sort of thing that interests me...the big downer to stuff like the Lily Drone is the lack of collision avoidance. From the spec of the DJI, you have something that could track a sports event without any expert intervention...if it lives up to the rather gushy video, film-makers will kill for it. The Active Track, combined with obstacle avoidance, could remove the need for helicopters to film, especially as the DJI has a decent top speed (for example, image the TDF, which normally has 6-8 camera bikes...instead you could have team/commissionaire cars with a drone on-board, then set it off, get really close footage, and then return to the car when done). All you really need is live broadcast capability, and that is already available
DJI have definitely set their stall up in the Action Camera market, rather than the flyable toy stall...as such I'll be interested to see what GoPro come up with when they finally announce their drone, as I suspect they are the real competition now. there were rumours that they were planning on a more serious/professional bit of kit, so perhaps it will be something with the HeroCast tech built in...
tested had this video of the phantom 4's tracking in action. Looked like it had some issues if it was more than walking pace.
Gopro are seeing their market being overrun by the Chinese knockoff action cams which while they're not as good image quality wise as the gopros are something like a 1/5th the price with all the same feature. They are really struggling at the moment so seeing this new market that is basically the logical extension of their product I'm surprised they've taken this long to get on board.
They did partner with us drone company 3drobotics on the solo drone that was marketed as the first "gopro compatible" drone what ever the hell that means.
I would think they're coming to this party so late that they'll have trouble but their name still has some weight to it. People like dji and parrot and even 3drobotics have spent several model generations working out how to provide the sort of flying camera platform and their machines are well made reliable and have a host of features. DJI have the high end basically sown up at the moment 3Drobotics are trying to get in there but they don't have the same recognition or the product portfolio that covers your home user all the way up to professional videographers. People like parrot have the low end more affordable but still capable with some interesting machines like the new disco flying wing.
Something that's been facinating me is how drones have been changing the way tv is made before you wanted a shot from height you needed a crane or a helicopter shots that costs money and both have limitations. A professional level drone can shoot stabilized 4k video it can maneuver in tight spaces and still provide the vertical flight for a fraction of the cost. The advanced models like the DJI's inspire one can be piloted by one person and have the camera independently controlled and offer a 4k down link for live video.
I think there are 2 markets for action cameras...the "hold my beer while I do something stupid" brigade, who don't really care about quality...then there is the more serious sports market, where the quality of the video means something. GoPro obviously want both, however their strength is in the second, as if (for example) you're a skier/snowboarder doing some nuts tricks, the last thing you want is a low fidelity video to edit together. If they can come up with a "follow me" drone which can
I reckon they could do very well. They are, however, looking like they are going to be a bit late to the market, but don't under-estimate the strength of the brand, I see a lot of GoPro's mounted on bikes, and they have a huge group of sponsored nutters producing footage for them in terms of advertising.
If there was a drone that could do 50-60km/h, had a reliable close-range follow mode, and had a 10-15 minute battery life in a form factor of about 1.5-2kg (the Session 4 camera weighs 75g), then I'd happily carry it on the bike on holidays, deploy it at the top of a climb, and get some blinding descent footage (something like that would have huge commercial value too, as for races like the TdF, the motorbike cameras cannot get close enough to the cyclists).
Indeed I've seen a lot of side by side comparisons of the chinese knock offs vs gopro and the quality is still with the gopro though some of the newer moderate priced knockoffs are getting a lot closer than they were. I've been looking at them as secondary cameras for racing quads and you have to factor into it the fact you might tank it into a tree or otherwise damage or lose it there the extra cost vs the risk doesn't seem to be worth that extra quality. If it's strapped to your head/bike I can see it's a different equation.
I think this phantom 4 is a step towards the action drone concept it's not really there yet plus it's designed to do a lot of other things in that one package I suspect like action cams action drones would be a niche market with specific models tailored to that activity.
It also seems like at the moment this tracking feature is more marketing buzz and gimmick than actually be useful given the one non dji controlled demo I've seen the obstacle avoidance seems to work pretty well but I've not seen it in a complex environment. It's still new so it will continue to develop.
A lot of other people are in the same game too I saw this today on a prototype autopilot system with avoidance. Still looks a little slow for real use but progressing.
There is this one, which operates at higher speeds (30mph/48kmh). There are also a few projects looking to shrink/cheapen Lidar sensors, which would also help here (for example, the DJI drone only has 60' obstacle avoidance to the front...360' would be far, far better).