Another day, another Pixhawk/IRIS+ compatible crowdfunding project! Meet Percepto, now live on Indiegogo:
Our mission is to create a valuable open-source ecosystem, not just a product, to harness the power of a strong developer community for the future of drone innovation.
Building community - Help us bring Percepto out of the lab
What's In The Box?
Percepto is a tiny black box that consists of a Tegra K1 based processing unit and a camera which easily connects to almost any drone, enabling users to create, upload and use existing apps from one centralized place, known as the Percepto Marketplace.
Included Software
Percepto App Marketplace
Percepto App Marketplace is meant to be the 1st marketplace designated for drone apps. Percepto's open source platform allows drone owners to reuse and share software driving down the cost of of developing new drone capabilities.
What's Unique?
Plug N Play Percepto
Rewards
Compatibility
Technical Specifications
Comments
@ Superwalloon weight without the shell is about 25-35 grams, the finger camera can act as an FPV camera. If using an APM, percepto can get all the sensors data and display it. You only need an RX for the down link, Percepto has a video output through the micro HDMI port.
To be honest that is what i had in mind when we designed it. PURE FUN.
Thanks for validating it for the rest of the team.
@ Kirill Vasiliev Thanks for the comment, Its TOTALLY open source.
Regarding the 180 degree lens i think there might be a few issues but that is for a more detailed discussion.
@Thomas Stanley-Jones Unfortunately the code is not ready to be published yet. Developing a new feature for percepto should be easy if you are familiar with Linux based development environments.
If you ever developed computer vision using openCV, integration into percepto should be seamless. If you have not, you can use our own computer vision algorithms through our API's with ease to create new Apps.
The pledge reached $31,083, more than 50% of their goal. Most of people looking for nice little Tegra board.
@ Thomas Stanley-Jones Currently Percepto supports up to 3 cameras. For your suggested configuration you can use 2 Perceptos integrated. We will re approach this idea when the hardware needed will be at reach. But you can start exploring the possibilities with one Percepto and let us know your insghts.
@ Andreas Our first apps are copter oriented but Percepto can be integrated in to any unmanned vehicle or robotics application.
@Jiro Hattori Development of computer vision applications on Percepto is done in a Linux environment with common libraries such as openCV and opanCL some of which are pre accelerated by NVIDIA. We believe that this is the broadest common ground for computer vision developers.
@ hotelzululima We will be shipping on oct 2015. Weight of Percepto in its box is 140 grams but if using just the PCB without the box it can be reduced to about 20 to 30 grams. That is if you are in to DIY and mounting it within the IRIS+ shell. We are in close discussions with NVIDA and the K1 is most suitable to be integrated in a product in our planned time frame.
I'm really interested in this, but I would like to see what the code of an App looks like. I didn't find anything online to see what it would take to develop an App for Percepto.
choices, choices..
while the processor for this is the K1(2014) and its an excellent processor.. the tegra X1(2015) is an octocore and gives you 256 cores.. and quite a bit more performance and is already being targeted at car automation etc..
the tegra K1 Jetson dev board was released last summer for about 192.00.. for price point comparisons no companion camera no flight adaptation etc..
and its seems that the K1 supports opencl which gets you to all that openCV loveliness..
hmm when can you ship??? how much does it weigh(IRIS + is overweight with a gimbal and gopro..)
parallela 16 core starting to lose its shine in comparison... although really wish they would consider the X1 vs K1
hzl
Used to make jokes in 1999 about selling a bag of silicon around which a community of open source developers would form to build millions of dollars of software around it.
@Ravi, So you will provide hardware, SDK and application platform by using Tegra K1. It is not clear the detail of application on your platform. The reason to asking is that Tegra K1 may have outstanding performance for image processing but I do not know how to use.
Have you already started developing something similar to the eBumper setup that we saw on DIYDrones recently? Instead of 6 sonar units, 6 cameras, for the purpose of collision avoidance with better definition and range. Maybe each sideways looking camera needs to be 2 cameras (a total of 10 cameras then, including one up and one down) to better 'see' the range of objects moving against the background. It sounds like this system is perfectly capable.
-
1
-
2
-
3
of 3 Next