Clean Contours or Surface From Drone Data

[SCrispV2HMC]

Hello, I flew a hillside and processed the images in Pix4D Matic and Survey. I then exported both a TIN (LandXML) and Contours (shp).

The TIN is extremely pocketed and not smooth as I would have hoped (see images 1 & 2). The 2nd photo shows my TIN data on the left with little pockets everywhere next to USGS LiDAR data on the right which is smooth as a baby bottom.

The contours are very erratic (see images 3 & 4). There will be 7 contour changes over a distance of 1 ft etc. Then when viewing the surface (Photo 4) it looks like a whole ton of layers stacked on top of each other. 

I have both Simplified and Smoothed surface to 100% of points removed, but not much changes. Is there a way I can say have no vertical changes of y height over x distance or something? Or will this have to be reprocessed with less detail in the drone program (Pix4D)? I want to have the surface detailed enough that I can pick out small drainage channels, but not to the degree that it creates this kind of havoc. 

Thank you

Contours from exported TIN via LandXMLPocketed surface on left next to LiDAR data from USGS on rightContours exported from Pix4D with many erratic changes in short distanceThe surface from the TIN with so many changes and layers

[Pointdump]

Hi Steven,
Your data is too dense.
ReCap will thin points, either by percentage or number of points, with >>>Intelligent Decimation<<<.
Dave

[SCrispV2HMC]

Unfortunately I don't have ReCap and I don't know if my employer will pay for it. Maybe I'll try the free trial, but would need a solution for the future. 

Thank you

[neilyj666]

@Pointdump I found the recap decimation was a bit confusing first time I used it as the percentage is how much data is to be retained rather than removed—as it is in the Surface Simplification in Civil 3D—I entered 85% and so my cloud hardly reduced in size..!!!

[Pointdump]

Hi Neil,
"...a bit confusing..."
Yup. I have a big problem with a lot of the terms in ReCap. Like the term "decimation", which means one-out-of-ten, not however much you key in. I think "thinning" would have been more appropriate.
But don't get me started. 🙂
Dave

[neilyj666]

@Pointdump totally agree