Hi,
Would a LiDAR device (like this one: https://www.robotshop.com/ca/en/ydlidar-x4-360-laser-scanner.html). Be able to detect a round object similar to a metal tube of about 1/4" in diameter from located at a distance of 5" to 48" from the LiDAR ?
Thanks for any answers or hints.
Hey,
Since most of our LIDARs use the metric system, we’ll have to do some conversions. Lucky for you, I’m Canadian and therefore can deal with both… 
Lets start with your requirements:
According to the datasheet (page 5) of the RB-Ydl-01 (found under Useful Links on the product page), we have the following information:
For your object to be detected, it needs to be hit by at least one beam from the sensor. To see if this works, lets use the worst case, which is at your maximum distance of 48" (~121.92 cm) and an angle resolution of 0.52°.
At such a distance and with this angle of resolution, the arc length between two distinct points would be ~0.4356" (~1.1065 cm). Since your object is 1/4" or 0.25" (~0.635 cm), it means it is possible your object will be completely skipped by the sensor on some rotations (but maybe not all of them). Reducing the scanning frequency can reduce the angle resolution down to 0.48°, which provides slightly better results (~0.40" or 1.02 cm) but still not enough to make it work reliably every scan.
Doing the reverse calculation (starting from arc length of 0.25" (~0.635 cm)) and an angular resolution of 0.52°, you get a maximum distance of 27.54" (~69.967 cm) before the arc becomes longer than your object: i.e.: the space between individual measurements.
On top of the angular resolution / distance / object size issue you also have the potential problem of the object surface. Since it is made of metal, you’ll need to make sure it is a kind of metal that provides a proper surface for IR signals of the frequency used by the sensor. If it is not, you would need to change the surface of the object so that it returns the signal properly. Since the object is round (assume circular) and possibly IR reflective (metal), it is possible it would cause issues by not returning a strong enough signal to the sensor (i.e.: returning elsewhere instead).
Since you most likely cannot increase the size of your object or reduce the distance (they are requirements, after all), you would need a sensor that either scans faster or scans slower, therefore reducing the angular resolution (i.e.: less distance between individual measurements). For an object that is 0.25" (~0.635 cm) or wider (to the sensor) and a distance of 48" (~121.92 cm), you would need an effective angular resolution of 0.2984° or less to ensure capture on every scan.
That being said, as mentioned above, it is very possible your object will show up in one scan (just not every single one). Therefore, depending on your speed of scanning requirements, this may be acceptable (or not).
As you most likely know already, we offer many options for [rotational LIDAR (https://www.robotshop.com/ca/en/rotational.html?dir=asc&order=price). Most of the lower cost ones have an angular resolution of ~1°, so the RB-Ydl-01 is actually the best in that price range. Things do get better at higher prices, such as with the RB-Rpk-07 which reports an angular resolution of 0.225° (datasheet, page 9) in Enhanced Mode when used indoors.
We hope this information helps! 
Sincerely,
WOW ! Now that’s an award winning post answer !
Thanks a lot.
No the object composition can not be changed. I mentioned metal but could be fiberglass or even wood.
The RB-Rpk-07 looks great but way outside the cost range for this application. I’ll probably get a RB-Ydl-01 (when it gets back in stock) and try it out. I don’t really mind if it does not detect on every pass. As long as I have a few reading to get it’s distance and angle from the sensor I’ll be able to extrapolate it’s X/Y coordinate.
Thanks again for the great and detailed analysis.
Hey,
Haha!
Just keep in mind that different surfaces will produce different results. Therefore, you may need to calibrate your results based on the material of the target. If the device capturing the data or the software performing data analysis of the scan results has a way of being told the material you should be able to implement something reliable (provided your selection of materials even has an impact at all!).
Indeed, it is most certainly more expensive.
I’m quite curious on how well it will work out for you. If you feel up to it, either post here again (this topic) or make a new one about it at a later time. Maybe with some pics too, if you can!
As for waiting until it is in stock, I highly recommend to place an order now so that you get one reserved for you during our next shipment. Since this is not a high volume item, there is a chance that by the time the products gets here/back in stock enough orders were placed that none would be available (i.e.: never show up as “in stock”)…
Yeah, sounds pretty good then. It really only is a problem if you need real-time data that is accurate (like collision detection while moving). If you can wait for dozens of readings without issue, then probability says you’ll get one or more readings of the object eventually!
No problem.
Good luck with your project!
Sincerely,
