Hi everyone. I started a new project. Would like some input in what do you want to see in an inexpensive LiDAR.
So, if I understand correctly, it will use the laser to project an array dots like the kinect does with an IR LED. Then, a camera/image sensor of some sort to view the array of dots and use software to determine distances?
Will it be infrared? Will it work outdoors? What is the field of view? What is the spacing of the dots?
I am very interested.
... what do you want to see in an inexpensive LiDAR.
Simple. A working solution with range, resolution and high enough refresh to be practical. :)
Correct. I will start with visible laser. Camera will have filters, so it should work outside. At least I'm aiming at it. FOV will be near 90x90 degrees, there will be apx 15x15 dots. I'm thinking ASCII output through USB of XYZ coordinates of each reflection in relation to 0,0,0 where camera is located.
Would 15 refreshes per second be acceptable?
15fps sounds good.
Would the beam be split into 255 dots through a lens?
All the specs sound good, I just need something that will work in the Mojave sun.
Wherever you add mechanical means - complexity increases. Price will too. I'd rather find intelligent means to detect and track each spot on static screen. At least for now.
I agree. I wonder how the FDA measures "eye safe" energy levels. If you have a powerful, non eye safe laser diffused into 255 dots that are at eye safe levels individually, is that OK?
I'm using class 3R laser 5 mW. They're ok to use since blink reflex will limit exposure even for straight 1 beam. And 225 dots are definitely safe.
Why not use a scanning system? I don't think you're going to get enough power splitting your beam like that and your resolution will be terrible.
Personally, I'd probably go with a stereo machine vision system since you're going to have the computing hardware onboard anyways.
I agree that resolution will be low, but it should be enough for obstacle avoidance, or low altitude determination. There are multiple ideas out there, but this is the one I'll be developing, since I also want to offer it as a product. Not everyone needs sophistication of a stereo vision, but everyone need reliable rangefinding.