如何对OpenCV calcHist()生成的直方图执行polyfit()拟合？

Hey there! Let's figure out what's going wrong here. That error is telling you exactly the problem—your x array and histr array don't have the same number of elements, which polyfit requires to work properly.

The Root of the Error #

Looking at your line where you define x:

x = np.linspace(0,histr.shape[0],1); # <== ERROR HERE

The third argument to np.linspace is the number of points you want. You've set it to 1, so x ends up being a single value. But histr has 32 elements (since you specified [32] bins in calcHist). That's why you're getting the "expected x and y to have same length" TypeError.

On top of that, cv.calcHist() returns a 2D array (shape (32, 1)), while polyfit expects a 1D array for the y-values.

Fixed Code #

Here's your code with the necessary adjustments to get the polynomial fitting working:

import numpy as np
import cv2 as cv
from matplotlib import pyplot as plt
import numpy.polynomial.polynomial as poly

img = cv.imread('SomeImage.jpg')
color = ('b','g','r')
for i,col in enumerate(color):
    histr = cv.calcHist([img],[i],None,[32],[0,256])
    plt.plot(histr, color = col, label=f'{col} original')
    plt.xlim([0,32])
    
    # Create x array matching the number of histogram bins (32 points)
    x = np.arange(histr.shape[0])  # Gives [0, 1, 2, ..., 31]
    
    # Flatten histr to 1D and run polynomial fit
    coefficients = poly.polyfit(x, histr.flatten(), 4)
    
    # Generate the fitted curve and plot it
    fitted_curve = poly.polyval(x, coefficients)
    plt.plot(x, fitted_curve, linestyle='--', color=col, label=f'{col} fitted')

plt.legend()
plt.show()

Key Fixes Explained #

• Matching x array length: np.arange(histr.shape[0]) creates an array of integers from 0 to 31, which perfectly matches the 32 bins in your histogram.
• Flattening the histogram: histr.flatten() converts the 2D output of calcHist() into a 1D array that polyfit can process.
• Optional fitted curve plot: I added code to generate and plot the fitted polynomial so you can visually compare it to the original histogram.

内容的提问来源于stack exchange，提问作者meaning-matters