color_frame intrinsics vs align intrinsics (are same ?)

[scizors]

D415 camera
code :python

Problem: color intrinsics and align intrinsics are same why?
#code color_intrinsics:
import pyrealsense2 as rs
import numpy as np
import cv2
pipeline = rs.pipeline()
config = rs.config()
config.enable_stream(rs.stream.depth, 1280, 720, rs.format.z16, 6)
config.enable_stream(rs.stream.color, 1280, 720, rs.format.bgr8, 6)
profile = pipeline.start(config)
frames = pipeline.wait_for_frames()
color_frame = frames.get_color_frame()
intrinsics=color_frame.profile.as_video_stream_profile().intrinsics
output color_intrinsics:
width: 1280, height: 720, ppx: 630.005, ppy: 376.119, fx: 925.769, fy: 925.476, model: 2, coeffs: [0, 0, 0, 0, 0]

#code:align intrinsics
import pyrealsense2 as rs
import numpy as np
import cv2
pipeline = rs.pipeline()
config = rs.config()
config.enable_stream(rs.stream.depth, 1280, 720, rs.format.z16, 6)
config.enable_stream(rs.stream.color, 1280, 720, rs.format.bgr8, 6)
profile = pipeline.start(config)
frames = pipeline.wait_for_frames()
align_to = rs.stream.color
align = rs.align(align_to)
aligned_frames = align.process(frames)
aligned_depth_frame = aligned_frames.get_depth_frame()
intrinsics = aligned_depth_frame.profile.as_video_stream_profile().intrinsics
output:align intrinsics
width: 1280, height: 720, ppx: 630.005, ppy: 376.119, fx: 925.769, fy: 925.476, model: 2, coeffs: [0, 0, 0, 0, 0]

[ev-mp]

@scizors hello.
The intrinsic (as well as the extrinsic) parameters for the aligned stream are borrowed from the stream the data is aligned to, i.e the color stream:

align_to = rs.stream.color
align = rs.align(align_to)

[scizors]

Hello,
Thanks for the reply .
-I have used checker-board to calibrate the RGB camera D415
-Then, I imported the images into Matlab finally I got the intrinsic values from Matlab
width: 1280, height: 720, ppx: 628.505325548172 , ppy: 375.139062149491 , fx: 908.237035106199 , fy: 908.710269684772

From intelrealsense
width: 1280, height: 720, ppx: 630.005, ppy: 376.119, fx: 925.769, fy: 925.476, model: 2, coeffs: [0, 0, 0, 0, 0]
camera is already well calibrated while buying it. why there is major difference in the focal length (fx and fy)? will this deviation affect the accuracy ?

which parameters should I use now ?

[ev-mp]

@scizors,
The focal length affects the object's magnification, so in case fx/fy are skewed the size of the objects would be scaled and it is quite straight-forward to devise a test that will measure ruler's size for verification.
A more methodical approach is to use the DepthQuality white-paper and the accompanying rs-depth-quality tool to characterize depth data and to see whether the metrics are within the specifications.

As for using 3rd part calibration with camera - currently this option is relevant only for OEM customers.
When it come to the usage then with the live camera you can only use the factory calibration (that has been recently extended with on-chip calibration feature)
The way you can try the alternative calibration is wrap the live device with software_device and have it calibration established with Matlab-produced data.

[ev-mp]

Assume resolved