I'm trying to develop a quality measurement (kind of a metric) for the reliability / quality of given point cloud data (coming from a 3D LiDAR) at a given point or region. It should come into account when there are situations where LiDAR data is not as usable as we want it to. Imagine a very foggy and/or rainy environment for example, where the rain/fog particles cause scattering.

So one approach I am thinking of is to observe the x- and y-position at the "position of interest" for a while, take the regarding z-values during this time-slot and compute the variance (considering the data is normally distributed).

Am I right in saying that if this variance exceeds some threshold, I could interpret this as "bad"/unreliable point cloud data at this point? Of course this only holds for static regions - if I have dynamically changing "regions/points of interest" obviously the variance will be no meaningful metric.

  • 1
    $\begingroup$ Welcome to Robotics Chris, but I'm afraid that questions like this are off-topic because there are many ways to solve any given design problem. We prefer practical, answerable questions based on actual problems that you face, so questions which ask for a list of approaches or an evaluation on a method are off-topic. Please take a look at How to Ask & tour for more information on how stack exchange works. $\endgroup$ – Mark Booth Mar 13 '19 at 14:51
  • $\begingroup$ @MarkBooth thanks for the hint and sorry for not being more specific and practical. I'll edit the question. $\endgroup$ – Chris Mar 14 '19 at 9:38
  • $\begingroup$ Adding a 'thanks in advance' section to a question is not required, we all tend to be thankful for the people helping us, and expect other people to be thankful too, so saying it just adds noise to the question and distracts people from the problem posed. It may seem counter intuitive, but excessive politeness can itself be impolite, as giving people extra text to read, even if they ignore it, is disrespectful of their time. $\endgroup$ – Mark Booth Mar 14 '19 at 17:39
  • $\begingroup$ I'm still not certain how practical an answer to your question could be. Normally we ask people to include details of what they want to achieve, what they tried, what they saw & what they expected to see, but this is more difficult when asking about theoretical solutions rather than solutions you've actyually tried. In any case, working through the Robotics question checklist might help in making your question clearer. $\endgroup$ – Mark Booth Mar 14 '19 at 17:47

Using your suggested method you can rule out the presence of moveing elements in the scanned area. I do no think you can rule out (slowly moving) fog. The metric will rule out moving objects/particles but there are other factors which influence readings wich cannot be ruled out.

Static measurments will mean that reflections do not change and any environmental conditions that effect the measurement (e.g. sunlight) will also be static. These factors can influence a reading and their influence remains constant in a static setting therefore you will not see any variance in the readings if the reading is affected by reflections of the laser or by sunlight.

| improve this answer | |
  • $\begingroup$ Thanks for attempting to answer this question @50k4, but I'm not sure this answers either the original question, or the revised one. Does the edit to the question make it more practical for you to answer this more fully? $\endgroup$ – Mark Booth Mar 14 '19 at 17:49

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.