I just want to get advice on how to clean GPS directory. Here is the scenario, though I can't give so much detail or data because of some rules.

I have my dataset timestamp, latitude, and longitude. When I graph it using graphing tools, there are many outliers on track. For example, there are track that has almost clean data, just small noise. But there are tracks that has duplicates, maybe, because the surveyor tends to get measurement twice or thrice.

My question is that do you know some algorithm for removing outliers and then smoothing. I had tried getting the bearing angle from one point to the next point, then filtered if bearing angle exceed some threshold. But still, there are many outliers remaining. As I observed, some points are on location which the surveyor is in resting state, but because it is not on track, it is considered outliers. Also, I have some research about Kalman filter but I can't apply it on my code. By the way, I am using Python here.

Thank you in advance for your support.


Browse other questions tagged or ask your own question.