Slashdot Mirror
Zero Zero's Camera Drone Could Be A Robot Command Center In The Future (ieee.org)
Tekla Perry writes: Zero Zero Robotics comes out of stealth today with the Hover Camera drone that uses face and body recognition to follow and photograph selected subjects. Company cofounder Meng Qiu Wang explains why he did the engineering in China (he built a team of 80 that worked two years on the project), and how this flying camera will evolve to be a navigation and control system for future home robots. According Zero Zero cofounder and CEO Meng Qiu Wang, "It has two cameras. The front viewing camera is a 13-megapixel camera that records video, but also has Simultaneous Localization and Mapping (SLAM), an algorithm that allows it to determine where it is. It also has a down-facing video camera, running an algorithm called optical flow, that looks at ground at 60 frames per second, so the Hover knows when it moves and can correct itself. These visual sensors are giving inputs and actual position and speed, meanwhile, the accelerometer and gyroscope gives relative position. All these signals are fed into the flight control algorithm, so when I throw it up in the air, it can just hover there." With a price of less than $600, it should compete well against the expensive DJI Phantom 4, which is already available on the market for $1400 and features autonomous flying and tracking features.
Very nice! These guys are really thinking about what they're doing.
a) Downward camera stabalization is circa 3 years ago, nowadays its done with two cameras in stereoscope.
b) The "follow" me feature on the Hover don't work, see the example video.
More to the point, watch the video, its awful, shaky. What makes this drone a toy is the lack of a camera gimble, so the image shakes as the drone flies making the footage useless. Its also likely why b) is happening. If you can't see a stable image, neither can the drone, so its tracking is terrible.
As to this marketing line, really get the product to work first. A neat little device like that would sell brilliantly as a smart camera drone, but not if the basic camera recording does not work!
This is just what we need, making it easier for drones to invade people's privacy. If you're flying high enough, small movements in the position of the drone won't have any effect on the quality of the picture or video. After all, this technology wasn't needed to get good photos and videos from helicopters. This is really for things like using drones to look in people's windows and spy on them. We don't need that.
That has to be the dumbest thing I have ever heard. He can throw it in the air and it can just hover there? It took you two years to build that with 80 people? Christ.
the robots are coming, they will take your jobs....
http://spectrum.ieee.org/view-from-the-valley/robotics/drones/camera-drone-could-be-a-robot-command-center Corrected link.
http://spectrum.ieee.org/view-...
09 F9 11 02 9D 74 E3 5B - D8 41 56 C5 63 56 88 C0 45 5F E1 04 22 CA 29 C4 93 3F 95 05 2B 79 2A B2
every moment she leaves her house can come true! Thank you China.
Just what high-end stalkers have been waiting for!
So, basically a flying optical mouse with long focus optics and pid correction? Will you sell it for 300 with shipping?
The only hard part about building a quad like this would be making all your software closed-source. The opencv libraries have loads of code to do SLAM and Optical Flow, and it's very basic computer vision/quad navigation stuff. You can do a setup like this for $300 and the only tricky bit is doing the "throw it up and it stays there" part, which certainly isn't dozens-of-engineers-for-two-years tricky. It's just something that requires properly tuned PID attitude controllers - like all quad stabilization code - and some quaternion-space orientation and displacement calculations. The 13MP camera is just for stills, as the SLAM will be downsampled (SLAM and Optical flow actually *benefit* from blurry images) to a computation-friendly size, for the most part.
GPS-independent position hold and navigation using sonar/lidar rangefinders, downard-facing optical flow cameras and forward-facing video cameras has been around for a few years and was commercially available for $300 in the Parrot AR.Drone 2 a while back.
All your base are belong to us.
Does it avoid trees wires and such?
The idea of a hover platform to provide guidance to ground-based robots would seem to make sense. Would be pretty noisy, tho.
"We receive as friendly that which agrees with, we resist with dislike that which opposes us" - Faraday
Flying cameras are drones designed for use by consumers that don’t want to learn how to fly a drone; they just want to take pictures....In a few weeks, I’ll be attending my son’s high school graduation in Silicon Valley, with, I’m sure, my view obscured by parents using pads and phones and selfie-sticks to record the moment. By next spring, I’m betting at least a few of the selfie-sticks and tripods are going to be replaced by camera drones.
Flying so near large groups of people is against the law.
Besides, a quadcopter will most probably crash, as the WiFi signal from the numerous smartphones of a large group will interfere with the radio command signal.
One must learn to fly anyway, otherwise it is just a recipe for a disaster.
Combine this with Facebook's face recognition, and your home can watch everyone as effectively as the sensors in "Minority Report". Of course then the hacking community/government will tap into them the way many do with the cameras in smartphones or IoT appliances and then...well. The Chinese government would probably LOVE one of these in every home. (Many of their products, one infamous router in particular, are know for existing back doors...not closed even after discovery).
"Imagination is more important than knowledge" - Einstein
foe example in slomo shots you can see all four propellers spinning anti clockwise :))
Who logs in to gdm? Not I, said the duck.