Would Nvidia support my project ?

Hello Nvidia,

i come here to ask you if you could give me some support and help to make me going further in my project.

I explain :

I have made a stand-alone sky survey system (beta version) :

  • home made 2 axis motorized mount
  • planetary camera (ZWO ASI178MC)
  • old but quite good Canon FD 50mm F/D 1.4 lens

Here is a pictures of the system :

First, i decided to use raspberry pi 3 B+ to control the mount and the camera, using Python and common libraries like Tkinter, Numpy, Pillow, OpenCV etc.

Everything works fine. The mount control is ok (i use an absolute gyroscope/compass to know the mount position).

The camera control is also ok but as i perform real time treatments of the camera acquisition, i need powerful CPU and raspberry pi is not the good choice for that kind of work.

Here is what my camera control software looks like :

External Media

It is not really user friendly but it works great.

So i decided to get a new SBC to control the camera (i keep the raspberry to control the mount). In order to get minimum software rewrite, i decided to get an Odroid N2 to control the camera. It works great.

But my project is a long term project. I need to improve the mount (i must make a new one, the first one allows me to validate some electronic and mechanical choices), i will have to get a new camera (ASI178MC is not the best choice for my project) and maybe i will have to find a new lens. All my budget is dedicated to the new camera (and maybe new lens).

As my project is a long term project, i would have like to test jetson Nano (but i can’t spend money for it ; as i said, my money left will be used for camera and lens which are really important pieces if my system).

Odroid N2 is quite good but i think maybe jetson Nano could be better for real time treatments and jetson Nano is really interesting SBC because :

  • Cuda
  • Nvidia support and experience (Odroid is far far away from Nvidia experience and support).

So, my request is quite simple :

Is Nvidia interested in my project and could support me giving me a Jetson Nano (and its alimentation) ? If the answer is yes, i could make some tests and try to get better results with Nano, using Nvidia experience and support (but i will take me some time because i will have to make a new version of my software and it is a lot of work).

I would be very pleased if Nvidia could support my project, really.

You can see some of my results on my facebook page (but you will have to scroll down to find pictures and videos) there :

[url]Redirecting...

Anyway, even if the answer is no, i appreciate a lot Nvidia work (i won’t give you the list of Nvidia video cards i have put in my computers but the list is quite long ; my first graphic card was a Riva TNT and the second was a GeForce 256 and so on).

Many thanks for your reply and have a nice day.

Alain

Hi Alain,

Would you be interested in a Jetson TX2 for this project?

-BG

Hello BG,

TX2 could be interesting, mainly with his GPU. But it is much more expensive than Nano. I doubt nvidia will give me a Nano but i am sure they won’t give me a TX2 unfortunately.

Alain

Hi Alain, thanks for sharing your cool project! We will get in touch with you over e-mail regarding your request. It would be great if you could add Nano to your camera rig and share the results.

Hello dusty_nv,

many thanks for your reply. I will wait for your email.

I must say i am very curious with Jetson Nano and more important, that kind of SBC ges all Nvidia team behind. That’s quite impressive.

Have a nice day.

Alain

Hello,

i hope you are still thinking positively about my demand.

Some small explanations about my system :

i decided to develop my own acquisition software because i do want to apply real time filters to my live videos. Astronomy cameras (mainly Sony sensors) are quite good but for deep sky imaging, you have to raise exposition time and camera gain. This means you get very noisy pictures (with heavy light pollution most of time).

With common acquisition software, the picture or video you get is RAW format without treatment. You make video or picture treatments after acquisition.

For my project, i want to make a kind of visual observation (real time) with a camera. That is to say i need to clean the pictures to get the best result i can get with the camera/lens so i had to make my own software to use the images filters i want.

But to do that, i need powerful SBC. That’s why jetson SBC could be interesting.

As i make captures with real time filters, CPU or GPU is under heavy load for hours. Considering jetson nano for example, this means :

  • i have to use 5VDC 4A power supply
  • i have to use an active fan to cool the board.

I hope those quick explanations will help you to decide wich nvidia system would be the more suitable to help me in this long term project.

When i say long term project :

  • my system works and i will develop it (mostly camera, lens, mount)
  • i will use it for years
  • concerning nvidia SBC, i need to learn how to use the GPU and i must say it’s quite complex so i will need some time to start to get interesting results but i think it worth the heavy job.

Best regards.

Alain

Hello Nvidia,

Well, finally, i guess you are not really ready to support my project. For sure, astronomy is not really the best subject to talk about jetson nano.

It could have been interesting to try to get some significant improvements of my software with nano and Nvidia GPU but in fact, i am not sure i would have succeed to manage Jetson Nano; i am not a Cuda specialist.

Anyway, Odroid N2 gets a very very good CPU big.little architecture and it seems works goes on to be able tu use Mali G52. So i will concentrate my efforts on Odroid N2. The results will be good.

Many thanks for all. Jetson nano seems to be a good sbc. It will allow people with few money to make interesting projects.

Bye bye.

Alain

If you know what big.little is and can build that thing, I think you’ll be fine without Nvidia :)

I haven’t read the whole thread but I assume you’re trying to track stars, align the images, and remove noise.

I am nearly certain there are tutorials out there already to do that you can adapt to Nano. You may also want to have a look at:

And

Yes, i know those things.

Thx.

Alain

Sorry for the delay, my teammate said they were going to get in touch and send you a unit, let me remind them.

Hello dusty_nv,

I am sorry, i thought you changed your mind.

There is no obligation at all. With our without support, i do like nvidia work and products.

Have a nice day.

Alain

Hello dusty_nv,

as you know, your teammate has contacted me. Everything is ok.

I would like to thank you and Nvidia to trust me and my project. It’s really cool.

I will learn a lot of things with Jetson Nano and i hope i will get very good results.

Nvidia rocks !!!

Alain

Hello dusty_nv,

you can follow our collaboration here :

[url]https://devtalk.nvidia.com/default/topic/1056064/jetson-projects/electronically-assisted-astronomy-with-a-jetson-nano/[/url]

Anyone is welcome !

Alain