Python crashes while calculating pointcloud

[Panepo]

Required Info
Camera Model	D400
Firmware Version	05.08.15.00
Operating System & Version	Win10 version 1709 (OS Build 16299)
Kernel Version (Linux Only)
SDK Version	2.9.0

Hi there,
I am trying to determine point cloud on python wrapper and find a problem. The python environment ALWAYS crashes if the resolution of the color and depth is not the same. It seems crash at the following position:

    # Tell pointcloud object to map to this color frame
    pc.map_to(color)

    # Generate the pointcloud and texture mappings
    points = pc.calculate(depth)

And there is a message pops on windows event viewer:

Faulting application name: python.exe, version: 3.6.3150.1013, time stamp: 0x59e338b7
Faulting module name: ucrtbase.dll, version: 10.0.16299.125, time stamp: 0x70f70cc4
Exception code: 0xc0000409
Fault offset: 0x000000000006b70e
Faulting process id: 0x381c
Faulting application start time: 0x01d3967a103204a0
Faulting application path: C:\ProgramData\Anaconda3\python.exe
Faulting module path: C:\WINDOWS\System32\ucrtbase.dll
Report Id: af310638-a0c9-4112-a5ed-bc9c05ec5a6e

[zivsha]

Hi @Panepo ,
I can't seem to reproduce this. Which camera are you using?
I've added the following lines to export_ply_example.py and it doesn't occur:

cfg.enable_stream(rs.stream.depth, -1, 640, 480, rs.format.z16, 30)
cfg.enable_stream(rs.stream.color, -1, 1920, 1080, rs.format.rgb8, 30)

Which streams were you using to get depth and color?

[Panepo]

Hi @zivsha
I have turned align on, maybe it is the problem?
I've modified export_ply_example.pyto this to fit my camera config, and this code crashes every time.

[zivsha]

Hi @Panepo ,
I get a crash as well when trying your code, the reason is that you are mapping the pointcloud texture before you calculate it. just reverse the two lines and it should work:

    points = pc.calculate(depth_frame) #First clculate
    pc.map_to(color_frame) #Then map texture

[zivsha]

I've marked as bug since it shouldn't crash like that.

[zivsha]

Does not reproduce with PR #1278 . Closing as fixed. @Panepo if you still see the issue please let us know or open a new issue