Python 3.6中无法导入OpenCV 2.4.13的cv2模块求助

Alright, let's work through this DLL load error—this is a super common snag when setting up OpenCV with Anaconda, but we can sort it out step by step.

First, I notice you might have mixed up your paths in the description: you said OpenCV is at C:\Users\harsh\Anaconda3 and Anaconda at C:\Users\harsh\opencv—that's almost certainly a typo. I'll proceed under the assumption that Anaconda lives at C:\Users\harsh\Anaconda3 and OpenCV is at C:\Users\harsh\opencv; if that's not the case, adjust the paths below to match your actual setup.

Here's what to try:

1. Fix Environment Variables for OpenCV DLLs #

OpenCV relies on its DLL files being accessible to the system, and this is often the root cause of the error:

• Navigate to your OpenCV build folder: for 2.4.13, the DLLs are in either C:\Users\harsh\opencv\build\x86\vc14\bin (32-bit) or C:\Users\harsh\opencv\build\x64\vc14\bin (64-bit). Pick the one that matches your Anaconda architecture (check if your Anaconda is 32/64-bit by running python -c "import sys; print(sys.maxsize > 2**32)" in Anaconda Prompt—True means 64-bit).
• Add this bin path to your system's Path environment variable:
  1. Right-click "This PC" → Properties → Advanced System Settings → Environment Variables.
  2. Under "System Variables", find Path, click "Edit".
  3. Click "New" and paste the full bin path you found.
  4. Restart your computer—environment variables don't take effect until you do this.

2. Verify Your cv2.pyd is Python 3.6 Compatible #

OpenCV 2.4.13's official Python bindings are typically built for Python 2.7, not 3.6. If you copied a cv2.pyd from the default OpenCV Python folder, it's almost certainly for Python 2.7, which won't work with your Anaconda 3.6 setup.

Instead of manually copying files, use Anaconda's package manager to install a compatible version—it handles all dependencies automatically:

1. Open Anaconda Prompt.
2. If you're using a specific environment, activate it first (e.g., conda activate base for the default environment).
3. Run this command to install OpenCV 2.4.13 (if available):

   conda install -c conda-forge opencv=2.4.13
   

   If that version isn't available for Python 3.6, install the latest compatible version with:

   conda install opencv
   

3. Validate the Fix #

After installing via conda (or fixing the environment variables and cv2.pyd):

1. Restart your IDE (like Spyder) or open a new Anaconda Prompt.
2. Run these commands to test:

   import cv2
   print(cv2.__version__)
   

If you don't get an error and see the correct version number, you're good to go!

Common Pitfalls to Avoid #

• Mismatched architectures: Never mix 32-bit OpenCV with 64-bit Anaconda (or vice versa)—this guarantees DLL errors.
• Manual file copying: Manually moving cv2.pyd is error-prone; using conda is the most reliable way to get a compatible build for your Python version.

内容的提问来源于stack exchange，提问作者Harsh Katakwar