Loading multiple .PTX but only one gets processed

Hi,
I programmed a small tool that captures 3D frames from Kinect V2 and saves .ptx files. It looks like they load correctly in MeshLab, and I think they import correctly in RC, but when I “Start” processing of the imported files, it looks like only a single file gets rendered in 3D. Should I expect RC to align and merge all the ptx, or is it normal that I can only calculate model based on a single .ptx file?

My tool generates a file header that puts the scanner registered position at 0 0 0, all the ptx files have the same header. Is this a proble, even if in RC I use “Unregistered”, “Color”, “Noise free”?

I have linked some of my .ptx files, each images is taken at a slightly different position.
http://www.zonetrigger.com/beta/ptx.zip
I’d really like to get some feedback if someone can get aligned…

Thanks!

Your PTX files are fine but have no RGB. All of them load in to RC fine try select all cameras and you will see it selects them all.

But with all your headers being the same will cause the problem, the point of a PTX file is that it needs the origin + xyz rotations to place it in the right position

After roughly re aligning all the ptx files this is what the headers should approxiamtly look like.

Note after the first 510 424 lines the next set have matrix changes

510
424
0 0 0
1 0 0
0 1 0
0 0 1
1 0 0 0
0 1 0 0
0 0 1 0
0 0 0 1
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
-1.482651 -1.218399 1.953995 0.499886 255 50 50
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
510
424
0.370113 0.089264 -0.092113
0.956057 0.280966 0.083739
-0.275030 0.958444 -0.075782
-0.101552 0.049421 0.993602
0.956057 0.280966 0.083739 0
-0.275030 0.958444 -0.075782 0
-0.101552 0.049421 0.993602 0
0.370113 0.089264 -0.092113 1
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
-2.598251 -2.114822 3.431961 0.499886 255 50 50
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
510
424
0.778123 0.189401 -0.135637
0.837288 0.280053 0.469594
-0.273040 0.958271 -0.084654
-0.473706 -0.057338 0.878815
0.837288 0.280053 0.469594 0
-0.273040 0.958271 -0.084654 0
-0.473706 -0.057338 0.878815 0
0.778123 0.189401 -0.135637 1
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
-2.065353 -1.689133 2.725082 0.499886 255 50 50
0 0 0 0.500000 0 0 0
510
424
1.285057 0.358051 0.220814
0.522863 0.234354 0.819568
-0.259555 0.959580 -0.108801
-0.811939 -0.155835 0.562557
0.522863 0.234354 0.819568 0
-0.259555 0.959580 -0.108801 0
-0.811939 -0.155835 0.562557 0
1.285057 0.358051 0.220814 1
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
510
424
1.065918 0.276746 -0.030699
0.724790 0.268516 0.634490
-0.270458 0.957890 -0.096428
-0.633665 -0.101713 0.766892
0.724790 0.268516 0.634490 0
-0.270458 0.957890 -0.096428 0
-0.633665 -0.101713 0.766892 0
1.065918 0.276746 -0.030699 1
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0
0 0 0 0.500000 0 0 0

Capture.JPG

Thank you, StevenRamsey!
I still have a few questions then…

Why do you say the PTX files have no RGB? I used dummy RGB data, with 0 0 0 for invalid points and 255 50 50 for valid points. Is this not ok?

My original col x row is 512x424, but data you posted shows 510x524. Is this a typo?

Did you align the clouds using a third party software then update the .ptx files with the new translation & rotation matrix?
Or did you use RC for the entire process?
Could you give me some details about the steps to achieve this?

Mine is 510 424 I have no idea where the 2 got lost.

When I ment no RGB I ment no real color, the PTX file you posted does have color just a single color for the whole file.

I did use a third party software (Cyclone) just to show you the changes in teh header information.

I don’t know much about Kinetic but to do what you are trying to do I think you will need to write your tool to capture the “Aligned” version of each frame if possible so that RC can see the header information. It will all depend on if the final model frmes are treated as individules and know there origin inside which ever software you are using.

It is a fairly manual proccess to register each individul frame, but douable, the data quality isn’t great so you may get strange results

Got it, I really appreciate the great info you shared, thanks a lot!