0
$\begingroup$

I've recently started working on a graphics/SLAM project. I've encountered a problem that requires the measurement or inference of the color / spectral reflectance (and ideally albedo) of points on objects in a cone-shaped field of view from ~1-10m away.

Currently, I'm using an IR sensor to determine distance and angle, and I'm also measuring reflectivity (over the wavelength of IR light that the sensor uses). Ideally I'd like to be able to assign a color value to every point of the point cloud generated by the IR sensor (the color measurement doesn't need to be particularly accurate).

I was thinking that maybe with the IR sensor I could estimate the dielectric function of the material and then subsequently infer the spectral reflectance from those measurements? But it's been several years since I studied optics/E&M and even then it was only at an undergrad level. If anyone could shed some light on the problem I'd be very grateful!

Thanks!

$\endgroup$
  • $\begingroup$ Just thinking offhand, can you think of any way to detect fluorescent materials which fluoresce in responds to a wavelength you cannot produce? My instinct says it should not be possible, though maybe there's some creative trick I'm not thinking of. $\endgroup$ – Cort Ammon - Reinstate Monica Aug 7 at 4:17
0
$\begingroup$

Using an IR sensor to detect color is not a recommended idea. The refractive index dispersion of many materials differs greatly between the infrared and visible spectrum due to complicated details about its electronic band structure, so you cannot infer the refractive index in the visible region by making measurements in the infrared region. You need to make measurements using visible wavelengths to determine the color.

$\endgroup$
0
$\begingroup$

dielectric properties are determined from doing spectroscopy (i.e. your IR source would have to be a tunable laser rather than a narrow band source and you would have to have some kind of synchronisation set up between source and receiver to determine wavelength).

Colour in the visible is dominated by entirely different processes than that in the IR. If you had a knowledge of the material at an extremely simplistic level than some level of inference would be possible, but otherwise forget it.

10 m away is not far and investing in an RGB camera is a small price in comparison to an existing LiDAR setup.

$\endgroup$
  • $\begingroup$ Thanks Paul, If I use an RGB camera, won't it be very difficult to assign to every point a specific color? I'm no physicist but it seems like at 10m it would be impossible to tell where the incident light is originating from $\endgroup$ – king_geedorah Aug 7 at 3:13
  • $\begingroup$ I work with LiDAR and we get aerial surveys done with LiDAR and high resolution imagery at 100s of metres. 10 m is no problem. The resolution of the point cloud and the image will likely not match, but you can interpolate. $\endgroup$ – Paul Childs Aug 7 at 4:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.