Python code to save depth images

[kongsgard]

Required Info
Camera Model	D435
Firmware Version	5.11.1.100
Operating System & Version	Ubuntu 18.10
Kernel Version (Linux Only)	4.18
Platform	PC
SDK Version	2.19.2
Language	Python

Issue Description

Hi! I have some issues with acquiring depth images from the camera. I've used the Python code included below (a slight modification of the provided example align-depth2color.py which also seems to give the same issue), but I only get images as follows:

As you can see, the normal RGB image is fine, while the depth image is very dark. However, in realsense-viewer the depth output looks good.
Any ideas on what could be the cause?

Code

import cv2
import imageio
import numpy as np
import pyrealsense2 as rs
import os
import shutil


def make_clean_folder(path_folder):
    if not os.path.exists(path_folder):
        os.makedirs(path_folder)
    else:
        user_input = input("%s not empty. Overwrite? (y/n) : " % path_folder)
        if user_input.lower() == "y":
            shutil.rmtree(path_folder)
            os.makedirs(path_folder)
        else:
            exit()


def record_rgbd():
    make_clean_folder("../data/realsense/")

    pipeline = rs.pipeline()

    config = rs.config()
    config.enable_stream(rs.stream.depth, 1280, 720, rs.format.z16, 30)
    config.enable_stream(rs.stream.color, 640, 480, rs.format.bgr8, 30)

    profile = pipeline.start(config)

    depth_sensor = profile.get_device().first_depth_sensor()
    depth_sensor.set_option(
        rs.option.visual_preset, 3
    )  # Set high accuracy for depth sensor
    depth_scale = depth_sensor.get_depth_scale()

    clipping_distance_in_meters = 1
    clipping_distance = clipping_distance_in_meters / depth_scale

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

    try:
        frames = pipeline.wait_for_frames()
        aligned_frames = align.process(frames)
        aligned_depth_frame = aligned_frames.get_depth_frame()
        color_frame = aligned_frames.get_color_frame()

        if not aligned_depth_frame or not color_frame:
            raise RuntimeError("Could not acquire depth or color frames.")

        depth_image = np.asanyarray(aligned_depth_frame.get_data())
        color_image = np.asanyarray(color_frame.get_data())

        grey_color = 153
        depth_image_3d = np.dstack(
            (depth_image, depth_image, depth_image)
        )  # Depth image is 1 channel, color is 3 channels
        bg_removed = np.where(
            (depth_image_3d > clipping_distance) | (depth_image_3d <= 0),
            grey_color,
            color_image,
        )

        color_image = color_image[..., ::-1]

        imageio.imwrite("../data/realsense/depth.png", depth_image)
        imageio.imwrite("../data/realsense/rgb.png", color_image)

    finally:
        pipeline.stop()

    return color_image, depth_image


if __name__ == "__main__":
    record_rgbd()

[RealSenseCustomerSupport]

---

HI kongsgard,

I tried your code and has no issue on saving depth image. It is just grey level image file and hard to save details.
But if you copy-paste the displaying part of code from opensv_viewer_example.py file (https://github.com/IntelRealSense/librealsense/blob/master/wrappers/python/examples/opencv_viewer_example.py), it would display out the results visually.

Thanks!

[RealSenseCustomerSupport]

---

Hi kongsgard,

There is another thing that you may try is replacing the line for getting depth_image with the following line:
depth_image = np.asanyarray(aligned_depth_frame.get_data())
=>
depth_image = cv2.applyColorMap(cv2.convertScaleAbs(aligned_depth_image), cv2.COLORMAP_RAINBOW)
This should give more color info to the depth image, and help show the depth image is good.

Thanks!

[vnt1537]

How can I save a depth image array created by depth image to csv or How can I retrieve and see the depth image information from .bag if I want to apply some algorithm to that data.
Also how would I normalise the depth image?
Thanks in advance!!

[RealSenseCustomerSupport]

---

Hi vnt1537,
You may refer to the example of read_bag_example.py:
https://github.com/IntelRealSense/librealsense/blob/master/wrappers/python/examples/read_bag_example.py

Hi kongsgard,
Any update from you? If nothing else is needed, this one will be closed.

Thanks!

[kongsgard]

Thank you for all the feedback!
This solved the issue for me.

[karlita101]

Hi kongsgard,

There is another thing that you may try is replacing the line for getting depth_image with the following line:
depth_image = np.asanyarray(aligned_depth_frame.get_data())
=>
depth_image = cv2.applyColorMap(cv2.convertScaleAbs(aligned_depth_image), cv2.COLORMAP_RAINBOW)
This should give more color info to the depth image, and help show the depth image is good.

Thanks!
While this gives a good visual representation, there would be a loss of depth information. Is there another way to keep the accuracy of the depth information in the png?
#8639