Edge Impulse video input source

kevin192291 · May 15, 2021, 8:19am

Hello!
First I want to thank the creators of this, I have been using OpenCV for quite some time and I find Edge Impulse a massive upgrade!

My question is how I can use input sources that aren’t the camera of my Raspberry Pi. I would love to run this on a server and get video from an IP Camera or Youtube video stream. These are easy things to do in Python and OpenCV, but I have no clue where to start with getting video from other sources in Edge Impulse.

If anyone could point me in the right direction I would really appreciate it!

Again, Thanks everyone!
-Kevin

dhruvsheth · May 15, 2021, 10:23am

Hi @kevin192291!
As far as I know, you can use the linux-python-sdk to alter python code used to run your object detection model. So, this script initiates camera here -

github.com

edgeimpulse/linux-sdk-python/blob/eb247d0ea02b8fc3a0adecd84d85e3fe9a1c6971/examples/image/classify.py#L20


runner = None
show_camera = False
def now():
    return round(time.time() * 1000)
def get_webcams():
    port_ids = []
    for port in range(5):
        print("Looking for a camera in port %s:" %port)
        camera = cv2.VideoCapture(port)
        if camera.isOpened():
            ret = camera.read()[0]
            if ret:
                backendName =camera.getBackendName()
                w = camera.get(3)
                h = camera.get(4)
                print("Camera %s (%s x %s) found in port %s " %(backendName,h,w, port))
                port_ids.append(port)
            camera.release()
    return port_ids

So, you could further modify this script to be using an input Video Source on Server, and the modelfile.eim hosted and accessible to the script. I’m not aware about the edge-impulse-linux library as a whole, but this script should mainly do your job.
Best, Dhruv

louis · December 15, 2021, 4:42pm

Hello @kevin192291,

For you information, we just created an example on our Linux Python SDK:

github.com

edgeimpulse/linux-sdk-python/blob/master/examples/image/classify-video.py

#!/usr/bin/env python

import device_patches       # Device specific patches for Jetson Nano (needs to be before importing cv2)

import cv2
import os
import time
import sys, getopt
import numpy as np
from edge_impulse_linux.image import ImageImpulseRunner

runner = None
# if you don't want to see a video preview, set this to False
show_camera = True
if (sys.platform == 'linux' and not os.environ.get('DISPLAY')):
    show_camera = False


def help():
    print('python classify-video.py <path_to_model.eim> <path_to_video.mp4>')

This file has been truncated. show original

You just need to download your modelfile.eim using Edge Impulse Linux CLI and do:

python3 classify-video.py <path_to_model.eim> <path_to_video.mp4>

Regards,

Louis

jmorris644 · March 16, 2022, 3:54pm

Is mp4 the only format allowed? I would be looking for H264.

louis · March 17, 2022, 8:31am

Hello @jmorris644 ,

I only tried with mp4 but I guess you could add some code to either convert it or to take H264 as an entry.
Let us know if you can make it work, that’d be useful for other people I guess

Regards,

Louis

MrAccelerated · July 22, 2022, 10:57am

Hi!
Is there any Windows version of this?

louis · July 22, 2022, 12:29pm

Hello @MrAccelerated,

I am not sure your question is related to this thread, maybe you’re looking for this: Is there are way to do classification for Video - #5 by janjongboom

Regards,

Louis

MrAccelerated · July 22, 2022, 1:30pm

@louis Yes! Thats what Im looking for. Thanks for your attention!

Kishenmjr · May 1, 2024, 1:02pm

how do you install device_patches
it says
pip install device_patches
Looking in indexes: Simple index, piwheels - Simple index
ERROR: Could not find a version that satisfies the requirement device_patches
ERROR: No matching distribution found for device_patches