Stream depth camera (Intel D455) with OpenCV

Python
,
1

Hello,

I’m trying to stream a depth video from Intel D455 Depth Camera via http to get it back in another machine.

I wanted to use cv2.VideoCapture(Input_Depth_Flux).

If in print(Input_Depth_Flux) i will get a matrix with every value for each pixels (which correspond to the distance), so i suppose i have to encode this or something but i’m a but lost how to do this…

Could you help me to get on the way

Here is the script I expected to adapt to stream depth, and when i put the depth_frame in argument you can see the error message i get:

Screenshot from 2022-08-19 11-19-24
Screenshot from 2022-08-19 11-19-241541×946 140 KB

I’m not advanced in python and opencv I saw some tutorial on this library but still not all clear, So feel free to explain even if it seems stupid for you :slight_smile:

2

Here is the script I’m trying to inspire from which display depth_stream via cv2.imshow

Screenshot from 2022-08-19 11-12-11
Screenshot from 2022-08-19 11-12-11569×903 31.4 KB

I precise it work perfectly and display the depth video

3

please post code, not images of it, thank you.

4

Well here is the code I use to get the image:


from realsense_camera import *
import cv2
import os

rs = RealsenseCamera()

while True:
    ret,bgr_frame, depth_frame = rs.get_frame_stream()
    distance = depth_frame[100,100]
    print(depth_frame)
    cv2.imshow("Depth Frame", depth_frame)
    key = cv2.waitKey(1)
    if key == 27:
            break

You can get realsense_camera module following this link :

The second code i try to adapt to stream depth is:

# Part 01 using opencv access webcam and transmit the video in HTML
import cv2
import pyshine as ps #  pip3 install pyshine==0.0.9
from realsense_camera import *

HTML="""
<html>

<body>
<center><img src="stream.mjpg" width='640' height='480' autoplay playsinline></center>
</body>
</html>
"""

StreamProps = ps.StreamProps
StreamProps.set_Page(StreamProps,HTML)
address = ('10.112.45.31',9000) # Enter your IP address 

rs = RealsenseCamera()
    


while True:


    
    StreamProps.set_Mode(StreamProps,'cv2')

    ret,bgr_frame, depth_frame = rs.get_frame_stream()
    #distance = depth_frame[100,100]
    #print(distance)
    #cv2.imshow("Depth Frame", depth_frame)
    
    print(depth_frame)
    Input_video =  depth_frame
    capture = cv2.VideoCapture(Input_video)
    print(capture.read())
    StreamProps.set_Capture(StreamProps,capture)
    StreamProps.set_Quality(StreamProps,90)
    server = ps.Streamer(address,StreamProps)
    print('Server started at','http://'+address[0]+':'+str(address[1]))
    server.serve_forever()

    key = cv2.waitKey(1)
    
    
    if key == 27:
            break
5

you’re using realsense_camera and something called pyshine (unmaintained, nobody uses it, doesn’t even have a working github repo). neither has anything to do with OpenCV.

your use of VideoCapture is nonsensical. you already have the depth data from the realsense library.

this is just a basic programming problem you’re asking about. your use of loops indicates that you have no idea what your code does. this forum is not for general programming questions.

6

thanks for helping it’s a bit more clear

effectively I don’t understand all my code I started python 3 weeks ago and I’m trying to improve by projects

Thanks for your encouragement