Jump to content
UltravioletPhotography

Hello from France


Ny Ando

Recommended Posts

Greetings !!

 

My name is Ny Ando and I am a PhD students in Electronic and Systems at the Institut Pascal in Clermont-Ferrand, France.

My PhD is about creating a semantic multispectral based embedded smartcamera, including multiple sensors : SWIR, Visible, Thermal and Ultraviolet.

With the lockdown in France and the lack of hardware, I retargetted my work on making some modification to a physically based renderer called Mitsuba2 in order to introduce the notion of multispectral imagery and to create a dataset with it. As I needed some insight from knowledgeable folks on UV light and matter interraction, I searched and found this forum.

My work has yet to produce tangible images as I am now blocked on multiple ideas.

For now in the mituba's work and mine, there are the following notions :

- light ray interraction with the surface, ie. ray direction, scattering, diffusion, or absorption;

- objects' spectral power diffusion and albedo;

- sensor quantum efficiency;

- camera filter;

- and lastly easy sensor quantum efficiency specification, and generally scence description.

And so now I am searching for the best way to represent UV light, and Infrared if possible, into the renderer.

Hopefully I will get back some feedback and I promise to show you the results once the whole thing works :smile:

Link to comment

Hi Ny! (Is that your first name?)

 

Your PhD project is interesting. Especially the SWIR part of it... but really the whole idea of combining images in bands so far away from each other is interesting. So you want to represent 4 channels in one image? If you only had three you could use the RGB channels, but with four you have to be more creative.

Link to comment

hi and welcome to UVP.

 

Your project seems to be ideal for our community as members' interest span the entire range from UV to SWIR.

Link to comment
Hi! Your interests overlap with mine in a lot of respects (particularly SWIR imaging, thermal imaging, UV imaging, and computational methods). For the last six months, I have been helping a student develop a semantic smart camera for thermal imagery, combining it with visible imagery using a Rasberry Pi and a FLIR module. We use the visible camera for tasks like detecting edges, then the thermal camera can determine the temperatures of various objects in the scene, and also understands things like "this surface is a stove."
Link to comment

Hi Ny! (Is that your first name?)

 

Your PhD project is interesting. Especially the SWIR part of it... but really the whole idea of combining images in bands so far away from each other is interesting. So you want to represent 4 channels in one image? If you only had three you could use the RGB channels, but with four you have to be more creative.

HI ! My first name is Ny Ando :smile:

Well the gist of the PhD is less about combining multiple imagery modalities into one image and more about augmented reality using "non-visible" features present in an image.

 

Hi! Your interests overlap with mine in a lot of respects (particularly SWIR imaging, thermal imaging, UV imaging, and computational methods). For the last six months, I have been helping a student develop a semantic smart camera for thermal imagery, combining it with visible imagery using a Rasberry Pi and a FLIR module. We use the visible camera for tasks like detecting edges, then the thermal camera can determine the temperatures of various objects in the scene, and also understands things like "this surface is a stove."

Hi ! Actually a colleague of mine works on a project using a raspberry Pi and a FLIR module, 3.5 if I am not mistaken, and segments an image using thermal data.

As for my project, the computing part will certainly be done on a NVIDIA Jetson Xavier as there will be a big Machine Learning component but I might be also using a FPGA for organising the data coming from the sensors.

And for my end goal, I think there will be a part of edge detection on the visible spectrum and then indication coming from the other bands.

On the algorithms' side, I have yet to implement anything but I based my research on remote sensing and particularly on a survey from 2019 on pansharpening methods. I don't know if I am going on the right track though.

 

hi and welcome to UVP.

 

Your project seems to be ideal for our community as members' interest span the entire range from UV to SWIR.

 

Hi and thank you for welcoming me :smile:

I hope I will be done with the renderer fast and then post some rendered images on here !!

Link to comment

Hi ! Actually a colleague of mine works on a project using a raspberry Pi and a FLIR module, 3.5 if I am not mistaken, and segments an image using thermal data.

As for my project, the computing part will certainly be done on a NVIDIA Jetson Xavier as there will be a big Machine Learning component but I might be also using a FPGA for organising the data coming from the sensors.

And for my end goal, I think there will be a part of edge detection on the visible spectrum and then indication coming from the other bands.

On the algorithms' side, I have yet to implement anything but I based my research on remote sensing and particularly on a survey from 2019 on pansharpening methods. I don't know if I am going on the right track though.

...your colleague is not named Benjamin by chance, is he? It would be funny if it was the same guy!

Link to comment
  • 2 months later...

Welcome to UVP, Ny Ando! You have a very fascinating PhD project and I wish you well on completing it.

 

It's difficult to imagine a camera with multiple sensors, but that is probably because I'm trying to envision 4 sensors in one of my Nikons!! And I know that is not quite the camera structure you are envisioning.

Link to comment

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...