Show Posts
1
General / Colorization of a TLS point cloud with TIFF 32bits float
« on: April 19, 2022, 11:31:30 AM »
Hello,
I work with thermal images via the multisystem approach. I would like to colorize a TLS point cloud with these to have a thermal aspect and a temperature value for each point.
I use RGB images (uint8) for the alignment (as master sensor), thermal images in color (uint8) and temperatures matrix (TIFF 32bits float) (both as slave sensors).
The differents sensors are calibrated and parameters of slave offsets are known and fixed.
Moreover, I georeferenced RGB images with GCP.
Then, I import the TLS point cloud with "import points" and I use "colorize dense cloud".
My problem is that I feel like only one image is used during the colorization process. You can see the colorization result on the attached screenshot.
In comparison, I try the same process with temperatures matrix as TIFF 16bits (pre-processed to find temperature later) and it works perfectly (all images are used).
So I would like to know if the problem of colorization comes from the 32bits float coding.
Thank and regards,
Vincent
I work with thermal images via the multisystem approach. I would like to colorize a TLS point cloud with these to have a thermal aspect and a temperature value for each point.
I use RGB images (uint8) for the alignment (as master sensor), thermal images in color (uint8) and temperatures matrix (TIFF 32bits float) (both as slave sensors).
The differents sensors are calibrated and parameters of slave offsets are known and fixed.
Moreover, I georeferenced RGB images with GCP.
Then, I import the TLS point cloud with "import points" and I use "colorize dense cloud".
My problem is that I feel like only one image is used during the colorization process. You can see the colorization result on the attached screenshot.
In comparison, I try the same process with temperatures matrix as TIFF 16bits (pre-processed to find temperature later) and it works perfectly (all images are used).
So I would like to know if the problem of colorization comes from the 32bits float coding.
Thank and regards,
Vincent
