3D corridinates of full Image from realsense Python interface or rs.format.XYZ32F format support

[ShettyHarapanahalli]

librealsense	2.22.0 RELEASE
OS	Linux
Name	Intel RealSense D435
Serial Number	745412070314
Firmware Version	05.10.03.00
Advanced Mode	YES
Usb Type Descriptor	3.2

I am Using real sense for a robotic application, I need to get the 3D coordinates of the image, I see the formate to get 3D data from the sensor is "rs.format.XYZ32F", but when I use this in python interface I get the below error -

AttributeError: type object 'pyrealsense2.pyrealsense2.format' has no attribute 'XYZ32F'

Can you please help me get 3D coordinates of the image, I am aware of "rs2_deproject_pixel_to_point" feature but I need to process 3D data from multiple points so the support format like XYZ32F or equivalent will be of good help.

[ShettyHarapanahalli]

Just wondering if we can do something like "Pixy2 CMUcam5" on any Realsense device, where we deploy ML model (instead of color-based filtering algorithm as in pixy), do processing on data and generate simpler/text outputs for further processing down the line.

[lramati]

As can be seen in the documentation the correct name for this enum value in python is rs.format.xyz32f. However, if you want a pointcloud from the D435, you'll need to enable the depth (and optionally color) and then pass these to the rs.pointcloud processing block