使用matplotlib的ArtistAnimation制作时序热力图遇问题求助

Hey there! Let's walk through fixing your animation code step by step. I see a few small issues that are likely causing the abnormal output you're seeing:

Key Problems in Your Original Code #

• Unnecessary shape parameter: plt.imshow() doesn't accept a shape argument—your data[:,:,t] is already a 34x34 array, so matplotlib will automatically detect its dimensions. Adding this parameter can throw off rendering logic.
• Unfixed color scale: Without setting consistent vmin and vmax for each frame, the color mapping will shift every time, making your animation jumpy or impossible to interpret consistently.
• Oversized figure: A 34x34 figure is way too large (this measurement is in inches!)—it'll hog memory and cause unnecessary rendering slowdowns or errors. A smaller, reasonable size (like 8x8) works perfectly, and you can adjust axes to fit the heatmap properly.
• Potential encoding gaps: When saving to MP4, you might need to explicitly specify a writer (like FFMpegWriter) to avoid encoding failures, especially if your system's default writer isn't set up correctly.

Corrected Code #

Here's the revised version with fixes and explanatory comments:

import matplotlib.pyplot as plt
from matplotlib.animation import ArtistAnimation, FFMpegWriter
import numpy as np

# Assume your data is a 34x34x100 numpy array; convert if needed: data = np.array(your_data)

# Use a reasonable figure size instead of the oversized 34x34
fig, ax = plt.subplots(figsize=(8, 8))

# Calculate global min/max to lock color scale across all frames
vmin = data.min()
vmax = data.max()

img_collection = []
for t in range(100):
    # Clear the axis to prevent overlapping frames from previous iterations
    ax.clear()
    # Plot heatmap with fixed color scale, no unnecessary shape parameter
    im = ax.imshow(data[:, :, t], interpolation='nearest', vmin=vmin, vmax=vmax)
    # Optional: Add a frame counter title for clarity
    ax.set_title(f"Frame {t+1}/100")
    # Append the image artist to your collection
    img_collection.append([im])

# Create the animation object
animation = ArtistAnimation(fig, img_collection, interval=200, repeat_delay=1000)

# Save with explicit writer (ensure FFmpeg is installed on your system)
writer = FFMpegWriter(fps=3, metadata={'title': '34x34 Heatmap Animation'})
animation.save('path/to/file.mp4', writer=writer)

# Optional: Preview the animation in a window
plt.show()

Additional Tips #

• Install FFmpeg: If you get writer-related errors, install FFmpeg (Windows: download from official site; macOS: brew install ffmpeg; Linux: sudo apt install ffmpeg).
• Test with smaller data: If issues persist, try a tiny test array (like 10x10x10) to rule out memory or data format problems.
• Consider FuncAnimation: For smoother, more efficient animations, swap ArtistAnimation for FuncAnimation—it's optimized for frame-by-frame updates. Here's a quick snippet for that:

  def update(frame):
      ax.clear()
      im = ax.imshow(data[:, :, frame], vmin=vmin, vmax=vmax, interpolation='nearest')
      ax.set_title(f"Frame {frame+1}/100")
      return im,
  
  animation = plt.animation.FuncAnimation(fig, update, frames=100, interval=200, blit=True)
  

内容的提问来源于stack exchange，提问作者Luna