Imageai: How to use ImageAI to detect real-time video stream
Hello!
Thanks to the great job.
How do I use ImageAI to detect target in real-time video stream?
For example, open the camera for real time detection
MRsJackYang
All 7 comments
Currently, ImageAI can only detect objects from a video file. However, we are currently developing the next version of ImageAI that will allow you to parse in a camera input and have detection performed on the live stream video. You will also be able to receive detection details in real-time. We are bringing in support for YOLOv3 in the next version. It will be available soon.
OlafenwaMoses
on 27 Jun 2018
Well, thank you very much. By the way, how long will the next version be? I can't wait.
MRsJackYang
on 29 Jun 2018
The latest version of ImageAI (v2.0.2) is now available and supports TinyYOLOv3 for fast and real-time object detection on Low-end devices like the NVIDIA Jets on GPU. Find links to documentation below.
http://imageai.readthedocs.io/en/latest/detection/index.html
OlafenwaMoses
on 16 Jul 2018
Also, support has been provided for device camera and IP camera video.
OlafenwaMoses
on 16 Jul 2018
Hello @OlafenwaMoses,
I am not able to view live detected video. Like, I can save the video in my directory but I wish to see the live detection. The input source is my webcam
from imageai.Detection.Custom import CustomVideoObjectDetection
import os
import cv2
execution_path = os.getcwd()
camera = cv2.VideoCapture(0)
detector = CustomVideoObjectDetection()
detector.setModelTypeAsYOLOv3()
detector.setModelPath("detection_model-ex-013--loss-0003.666.h5")
detector.setJsonPath("json/detection_config.json")
detector.loadModel()
video_path=detector.detectObjectsFromVideo(camera_input=camera,
output_file_path=os.path.join(execution_path, "holo2-detected3"),
frames_per_second=16,
minimum_percentage_probability=40,
log_progress=True)
print(video_path)
codebugged
on 8 May 2020
so is it possible to view the live video while it is processing ?
Alnasser0
on 15 May 2020
No @TaiTan0
Please ask @OlafenwaMoses to help us out
codebugged
on 16 May 2020
Related issues
c7jie
路
4Comments
bestbugwriter
路
5Comments
manuel1507
路
4Comments
ShobhaShiva
路
3Comments
pytesnim
路
5Comments
Most helpful comment
Currently, ImageAI can only detect objects from a video file. However, we are currently developing the next version of ImageAI that will allow you to parse in a camera input and have detection performed on the live stream video. You will also be able to receive detection details in real-time. We are bringing in support for YOLOv3 in the next version. It will be available soon.