基于Python Tkinter的GUI多视频展示：实时摄像头双画面开发需求

Got it, I’ve tackled this exact scenario before! Here’s a complete implementation that uses Tkinter for the GUI, OpenCV for video capture, and PIL to bridge the gap between OpenCV’s image format and Tkinter’s display requirements.

Step-by-Step Explanation & Code #

First, make sure you have the required packages installed:

pip install opencv-python pillow

Here’s the full working code:

import tkinter as tk
import cv2
from PIL import Image, ImageTk

class DualVideoStreamApp(tk.Tk):
    def __init__(self):
        super().__init__()
        self.title("Synchronized Camera & Grayscale Feed")
        
        # Initialize video capture (0 = default camera, change if needed)
        self.cap = cv2.VideoCapture(0)
        
        # Create two side-by-side frames
        self.frame_original = tk.Frame(self, padx=10, pady=10)
        self.frame_original.grid(row=0, column=0)
        
        self.frame_grayscale = tk.Frame(self, padx=10, pady=10)
        self.frame_grayscale.grid(row=0, column=1)
        
        # Labels to display the video streams
        self.label_original = tk.Label(self.frame_original)
        self.label_original.pack()
        
        self.label_grayscale = tk.Label(self.frame_grayscale)
        self.label_grayscale.pack()
        
        # Start the update loop
        self.update_frames()
        
        # Handle window close properly
        self.protocol("WM_DELETE_WINDOW", self.on_closing)

    def update_frames(self):
        # Read a frame from the camera
        ret, frame = self.cap.read()
        if ret:
            # Convert OpenCV's BGR format to RGB (required for PIL)
            rgb_frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
            
            # Create grayscale version
            gray_frame = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
            # Convert grayscale to RGB (since PIL needs 3 channels for PhotoImage)
            gray_rgb = cv2.cvtColor(gray_frame, cv2.COLOR_GRAY2RGB)
            
            # Convert to PIL Image objects
            pil_original = Image.fromarray(rgb_frame)
            pil_gray = Image.fromarray(gray_rgb)
            
            # Convert to Tkinter PhotoImage
            self.photo_original = ImageTk.PhotoImage(image=pil_original)
            self.photo_gray = ImageTk.PhotoImage(image=pil_gray)
            
            # Update the labels with new frames
            self.label_original.config(image=self.photo_original)
            self.label_grayscale.config(image=self.photo_gray)
        
        # Schedule the next update (30ms = ~30fps)
        self.after(30, self.update_frames)

    def on_closing(self):
        # Release the camera resource
        self.cap.release()
        self.destroy()

if __name__ == "__main__":
    app = DualVideoStreamApp()
    app.mainloop()

Key Details: #

• Synchronization: We read a single frame from the camera each time, then derive both the original and grayscale versions from it. This ensures they’re perfectly in sync.
• Format Conversion: OpenCV captures frames in BGR format, but Tkinter requires RGB. PIL handles this conversion seamlessly.
• Garbage Collection: We store the PhotoImage objects as instance variables (self.photo_original, self.photo_gray) to prevent Tkinter from discarding them immediately.
• Frame Rate: The after(30, self.update_frames) call updates the frames every 30 milliseconds, which gives a smooth ~30fps stream. Adjust this value if you need faster/slower updates.
• Cleanup: The on_closing method properly releases the camera when the window is closed, preventing resource leaks.

Troubleshooting Tips: #

• If the camera doesn’t work, try changing the index in cv2.VideoCapture(0) to 1, 2, etc., if you have multiple cameras connected.
• Ensure your camera is not being used by another application.
• If you get a "no module named" error, double-check that you installed opencv-python and pillow correctly.

内容的提问来源于stack exchange，提问作者Esha