Use a notch filter in front of camera
I have the camera Intel Realsense D405. I used it at the same time with a laser and I don't want to see the laser so I bought a notch filter (as explained here https://dev.intelrealsense.com/docs/optical-filters-for-intel-realsense-depth-cameras-d400). However, I wonder where to put the filter ? Do I have to dissamble the camera ? Indeed, if I just put the filter on the glass which is on the lens, I obtain the reflection of the lens on the filter in my images ...How to use those kind of filter with des D400s camera ?
-
Hi Cuau Lenaic You could explore the possibility of adding an anti-reflective coating to both sides of the notch filter.
https://en.wikipedia.org/wiki/Anti-reflective_coating
Anti-reflective coating is recommended when adding a transparent window in front of a RealSense camera, like the example in the image below.
-
Hi MartyG, thank you for answering. I am afraid that with my notch filter it is not possible but thanks for the idea !
-
If your laser operates in the near-infrared spectrum then the D405 would likely not be able to see it. This is because the D405's left and right infrared sensors are equipped with an IR Cut filter that makes only lower-frequency visible light perceptible to the sensors. Because the D405 is not equipped with the infrared projector component that other models in the 400 Series range have, it does not need to see IR light.
-
Unfortunately my laser operates at 660 nm and the camera does see it, which disturbs me for my application, that is why I was looking for a notch filter. But even if the notch filter works very well for filtering 660 nm (and I do not see the laser now as I wanted) it is reflective and interfere with green/red light making cyan and pink artifacts in my image, and reflectance as you can see when placing it on the camera's lens ...
-
Yes, 660 nm will be within the visible light range and so be perceptible to the camera.
How visible is the laser if you use the alternative RGB mode color from left infrared sensor instead of normal RGB? You can test this in the Viewer by only having Infrared selected and setting the format as RGB8 or BGR8 instead of the default Y8.
-
It looks like my previous comment was not sent (not sure but I try again : )
Haha yes I guess it is ressembling to USS Voyager ! But is is just a dome in colored silicone with white powder for stereovision to work better !I am not sure of the impact of the FPS in the visibily of the laser in my images ? I tested, and it did not change the aspect of the spot. For information i use this laser (https://www.acuitylaser.com/product/laser-sensors/short-range-sensors/ar100-super-compact-laser-triangulation-sensor/) which is not blinking
-
You mean between the spot and the camera (position 2 in the following scheme) ?
My filter is a little circle of diameter 1.2 cm. So I think in that case it will be hard to hide the laser from left and right camera, and doing that I see the filter in my images making a blue-ish circle in the middle.
I didn't test the depth quality in that case but I guess it will be poor and I will have reflection of ambient light on my filter ?
Is there a possibility INSIDE the camera D405 to place a filter ?
-
I was thinking closer to position 3, the point at which the laser exits the device (like putting a filter over the tip of a laser-pointer pen).
The depth image would likely have no depth data in the area of the image where the spot is because the intensity of the light will make it difficult for the camera to read depth information from that part of the object's surface. Though you could try applying a post-processing Hole Filling filter to see whether it can fill in an empty area on the depth image accurately. You can enable the Hole Filling filter in the Stereo Module > Post Processing section of the RealSense Viewer's options side-panel.
The available space inside the D405 casing behind the glass may be too small to insert a filter. It is possible to open the casing and use the camera caseless with its two internal boards (D401 Depth Module and Vision Processor D4 V4) but opening the case would likely invalidate the camera's sales warranty.
If the problem is glare from a reflection then a different kind of filter placed over the front of the camera may be better able to deal with the spot. A linear polarization filter negates glare on RGB images, as demonstrated in section '4.4 When to use polarizers and waveplates' of the optical filter paper, where a linear polarization filter makes a pane of glass much clearer.
Because almost any polarization filter will work as long as it is linear, they are inexpensive to purchase. You can search stores such as Amazon for the term linear polarizing filter sheet.
-
For position 3 it is not possible because if I do this the laser could not make its measure which is important for me.
Yes for now I am doing hole filling , but as the spot is a bit big I miss a lot of information in that spot that is why I wanted to hide it from the camera and be able to see behind.
Thank you for the idea of polarization filter. I have already tested it and it did not fix my problem unfortunately
-
Here is an example of what I mean when I talk about "laser disturbance", I observed it some days ago (2 images down). However, today, for a reason I cannot explain in the realsense viewer the surface seems ok in the red spot. Maybe it is the influence of the baby powder which was not here in the first example
-
The red spot reminds me of shining a colored torch-light on a wall. Even if the wall has a non-shiny texture that does not reflect well, a colored light is still going to show much more strongly on the wall than a normal 'white light'.
Assuming that the intensity of the laser light is proportional to the distance of the laser triangulation sensor from the surface, is there the option of moving the triangulation sensor further away from the dome to try to weaken the brightness of the spot?
-
Looking at the user manual for your triangulation sensor, it appears that the intensity of the reflected light is vital to calculating the measurement and so reducing the intensity would not be advisable. That brings us back to filtering the light out on the camera instead.
In a situation where you have an area of a surface that the camera cannot read and so it has no depth, casting stronger illumination on that area may help to bring out depth detail. A torch might be sufficient to test this possibility.
Please sign in to leave a comment.
Comments
21 comments