M3C2-significant change

[Maryem]

Hello CloudCompare Community!

First of all, thanks a lot for this wonderful software. I've been working on CloudCompare for my internship to compare the distance result giving by C2C, C2M and M3C2. The purpose is to compare 2 point clouds of a riverbank at 2 different dates. Everything worked perfectly with C2C and C2M. For M3C2, I followed all the instructions and tried to compute the distance for D=0.5m and d=0.1m then for D=0.5m and d=0.4m. The distance is correctly computed but all my significant change are equal to 1. I've been trying to test different values of D and d but I got the same results..

[daniel]

Do you mean that all the points are flagged as 'significant change'? I guess it's normal if you don't input a 'registration error'.

[pkegels]

Dear Daniel,

I have 2 point clouds: 1 with the DPI-7 scanner and the other is one with de Leica C10. I want to integrate those 2 clouds and I also
want to know what the accuracy is between those 2. The question is "can I integrate a point cloud from a handheld scanner in a point cloud from a TLS?" Is there a tool that i can use?

Can you please help me

Thx

[daniel]

Hi,

You can indeed 'register' the clouds with the 'Tools > Registration' methods.
http://www.cloudcompare.org/doc/wiki/in ... gistration

I believe the best solution for you would be to use the 'Point-pair based alignment' tool (see http://www.cloudcompare.org/doc/wiki/in ... itle=Align).

[Dimitri]

Hello CloudCompare Community!

First of all, thanks a lot for this wonderful software. I've been working on CloudCompare for my internship to compare the distance result giving by C2C, C2M and M3C2. The purpose is to compare 2 point clouds of a riverbank at 2 different dates. Everything worked perfectly with C2C and C2M. For M3C2, I followed all the instructions and tried to compute the distance for D=0.5m and d=0.1m then for D=0.5m and d=0.4m. The distance is correctly computed but all my significant change are equal to 1. I've been trying to test different values of D and d but I got the same results..

Hi Maryem,

a bit late, but I didn't spot your message. If all your points have a significant change = 1, this means, that the distance measured by M3C2 is significantly larger (at the 95 % confidence interval) than a measure of the roughness of the river bank and point density (see Lague et al., 2013 for more details).

Regarding the registration error: it is added to the confidence interval and used to evaluate the statistical significance of the change. But if you don't specify one, it is assumed to be equal to 0, and you could well have non significant change if the surface is very rough compared to the mean distance measured between the two clouds.

Finally, we do not compute a confidence interval when there are not enough points in the "projection cylinder" (typically 4 by default). If these cases occur, the significant flag should be = 0.

Dimitri