Google Releases Open Source 'Cartographer' (betanews.com) 26
BrianFagioli quotes a report from BetaNews: Machine learning and vision are essential technologies for the advancement of robotics. When sensors come together, they can enable a computer or robot to collect data and images in real-time. A good example of this technology in real-world use is the latest Roomba vacuums. As the robot cleans your dirty floor, it is using sensors combined with a camera to map your home. Today, Google releases Cartographer -- an open source project that developers can use for many things, such as robots and self-driving cars. "We are happy to announce the open source release of Cartographer, a real-time simultaneous localization and mapping (SLAM) library in 2D and 3D with ROS support. SLAM is an essential component of autonomous platforms such as self driving cars, automated forklifts in warehouses, robotic vacuum cleaners, and UAVs," says Google in a blog post. "Our focus is on advancing and democratizing SLAM as a technology. Currently, Cartographer is heavily focused on LIDAR SLAM. Through continued development and community contributions, we hope to add both support for more sensors and platforms as well as new features, such as lifelong mapping and localizing in a pre-existing map."
Mapping my house (Score:3, Insightful)
And phoning home, getting hacked because of shoddy security, 3rd parties using my data to schedule break-ins, and Google pimping my data without compensation. No thanks. I need IoT in my life like I need more bankers and lawyers.
Re: (Score:2)
"A Zionist Corporation having your info is bad if your a white male."
Thanks for at least not knowing the difference between "your" and "you're", like most racists, it made my day.
Re: (Score:2)
Thanks to the world I grew up in, I have a paranoid twinge every time I tell my cloud based smart thermostat that I'm going away on vacation for more than a few hours... Once in awhile, it's good to remember that in this world, that same cloud keeps high definition video records of everyone/everything that approaches my house, and especially enters it - and facial recognition software/databases means that a repeat offender won't be repeating for long.
Sure, there are people with access to my data who could
Re: Mapping my house (Score:1)
Hence why I will always prefer my homegrown solutions over mass marketed products.
Re: (Score:3)
It also makes it available. If google hadn't released it as open source, companies which might have needed such a functionality would have been required to roll their own. This would have been a terrible waste of money and time, as google already has a perfectly working implementation. Now they have resources free to do different things, e.g. to improve the existing implementation, or to focus some other feature of their products.
Re: (Score:3)
And they are contributing back to the community. Now YOU can build an empire on the back of work open sourced by google.
Documentation seems lacking (Score:2)
The only documentation [readthedocs.io] I can find gives some bare bones installation instructions, then refers back to itself, claiming to be the "complete documentation." What am I missing?
Re: (Score:1)
https://google-cartographer-ros.readthedocs.io/en/latest/
Re: (Score:2)
The rc/quad would most likely lack sensors accurate enough to record/merge/filter and build accurate results. The onboard FC of most systems, especially DJI aren't powerful enough to process SLAM. Likely you would need to have lidar/pulsed light/projection mapping etc on board with a powerful enough companion PC to do SFM/SLAM or photogrammetry.
Way to go iRobot! (Score:2)
You finally have what Neato Robotics have been doing since 2010.
But I guess you're cleverer than them -- you charge literally twice as much for a bot that does it.