GPU-Computing for Vision Recognition Platform as IoT

Hi my dear readers, today I would like to tell you three technical stories. It will be stories with a happy ending… and you know me..? I am not sure if that story happens exactly this way, but the story is true.

IoT Idea Story ;-).
IoT First of all, you probably know my passion for IoT? And also, as a tool builder, sorry to say, I do not like to buy the software but make it… some time ago, I made Multi Service Bus Core 2.0, and you can find this early version on the Downloads page of this blog. From that time the all idea involved a lot. And thanks to .NET Core 2.0+, I can use any Cloud provider, so I am making a cloud-agnostic solution like they say. So only any OS that supports .NET Core 2.0+ or Service that can handle requests with this technology is necessary to make a broker pattern. Many people think that sorry to say, IoT needs Service Bus… well, there never be more wrong :-). The only need is a well-secured, serialization optimized, and distributable ready communication. I would say; also, no central point, as a weak point, was not designed. All that I put in place.

HPC I can have ;-).
MINI HPC That is my server room, for the vision recognition, those 4 beauties, from AMD… well actually from eBay, once AMD sells only in Germany, my suspicious, and I was never lucky enough to buy from official AMD Store the GPU… but I have them and each GPU that class can easily compute faster than 100 CPUs… that is like 400 CPUs in super-secret place… only power supply and good Internet link are necessary. Going back to the subject, I have a solid server room 😛 and a super optimized, well-secured communication channel… what can I do with this? Well, we can monitor anything we need. We may use IoT without GPU and contact our server room :). to interpret the image, to store some amount of images… in short to do what we need.

GPU-less device ;-).
remote-home-monitoring Those are two devices, the BeagleBoard X15. Please do not ask how hard it was to but with the power supply :P. And the second is Logitech C922 camera. It must be a full hardware camera, not a camera sensor-like device. So now all that comes into conclusion that as iStorm.io platform I was able to make… well, a bit futuristic platform.

iStorm.io PoCs Happy Ending! Sells, for integration, open if you are interested in, please use a contact form. And I have just started… stay tuned, it is much more to come… ;-).
istorm-io

p ;).

Leave a Reply

Your email address will not be published. Required fields are marked *

*

This site uses Akismet to reduce spam. Learn how your comment data is processed.