How can I improve the accuracy of pyrealsense2 depth image

[LJu_B]

Hello community!

I'm recently study about image processing and i use Intel real sense Depth Camera D415.

And I find there are two ways to use depth camera. One is using intel real sense viewer program and another is using pyrelasense at python.

When I use viewer program, I get satisfactory result.

But when i test the code made by pyrelasense, the result is not.

So, how can I improve the performance of pyrelasense as much as intel real sense viewer's.

This is the test code that i use.

https://github.com/IntelRealSense/librealsense/blob/master/wrappers/python/examples/opencv_viewer_example.py

This is the result made by intel real sens viewer.

This is the result made by test python code using pyrlealsense

[MartyG · Accepted solution]

The sample that you linked to is a simple program for viewing camera data with OpenCV. and Numpty (an 'OpenCV Viewer', as described by the sample program's title, rather than an equivalent to the much more complex RealSense Viewer program).

You may be able to achieve depth image results more comparable to those provided by the RealSense Viewer by using Intel's 'distance to object' Python tutorial.

https://github.com/IntelRealSense/librealsense/blob/jupyter/notebooks/distance_to_object.ipynb

[LJu_B]

Thank you for your reply!

Thank to the code you recommend, I can know the difference of these

They suggest other method to make color map image!

And what I wanted was the method you suggested it.

Thank you again for your answer and it was very helpful.