nnspike.unit.webcam_video_stream
Classes
|
A threaded video stream reader for webcams with optional video recording. |
- class nnspike.unit.webcam_video_stream.WebcamVideoStream(src, save_video, save_path='', resolution=(640, 320), fps=30)[source]
A threaded video stream reader for webcams with optional video recording.
This class provides a non-blocking way to read frames from a video source by running the capture loop in a separate thread. It supports configurable resolution, frame rate, and optional video recording to file.
The threaded approach helps prevent frame drops and provides smoother video processing by maintaining a minimal buffer size and continuous frame updates in the background.
- Parameters:
src (
int|str) – Video source - can be camera index (int) or video file path (str)save_video (
bool) – Whether to save captured video to filesave_path (
str) – Path for saving video file (required if save_video is True)resolution (
tuple) – Video resolution as (width, height) tuple. Default: (640, 320)fps (
int) – Frames per second for capture and recording. Default: 30
Example
>>> stream = WebcamVideoStream(src=0, save_video=False) >>> stream.start() >>> grabbed, frame = stream.read() >>> stream.stop()
- start()[source]
Start the background thread for reading video frames.
- Return type:
- Returns:
Self reference for method chaining.
- update()[source]
Continuously update frames from the video stream in a background thread.
This method runs in a loop until stopped, constantly reading new frames from the video source to keep the frame buffer current.
- Return type:
None
- read()[source]
Read the most recently captured frame.
If video recording is enabled, also writes the frame to the output file.
- Return type:
tuple[bool, np.ndarray | None]
- Returns:
Tuple of (success_flag, frame) where success_flag indicates if the frame was successfully captured and frame is the image data as a numpy array, or None if capture failed.