numpy 数值微分

Question

我已经沿着数组的宽度绘制了一条曲线，即数组列的总和：

import numpy as np
from astropy.io import fits
import matplotlib.pyplot as plt

#def plot_slope(x, y):
#    xs = x[1:] - x[:-1]
#    ys = y[1:] - y[:-1]
#    plt.plot(x[1:], ys / xs)

hdulist = fits.open('w1_subtracted_2_deg.fits')
nodisk_data, nodisk_header = hdulist[0].data, hdulist[0].header

x = range(nodisk_data.shape[1])
y = np.sum(nodisk_data, axis = 0)
xticks = [0, 183, 365, 547, 729, 912, 1095, 1277, 1459]
long_marks = [24, 18, 12, 6, 0, 354, 348, 342, 335]

ax = plt.axes()
ax.set_xticks(xticks)
ax.set_xticklabels(long_marks)

plt.plot(x, y, linewidth = 0.5, color = 'red')
#plot_slope(x, y)
plt.title('Longitudinal sum of flux density per steradian')
plt.xlabel(r'Galactic longitude, $\ell$')
plt.ylabel(r'Summed flux density per steradian, $MJ.sr^{-1}$')
plt.grid(True)
plt.show()

plt.savefig('add_cols.png')

hdulist.close()

注释掉的函数是我尝试用数值计算曲线的导数，但我得到了一个

  File "C:/Users/Jeremy/Dropbox/Astro480/NEOWISE/add_cols.py", line 6, in plot_slope
    xs = x[1:] - x[:-1]

TypeError: unsupported operand type(s) for -: 'range' and 'range'

我发现的问题 here and here aren't quite what I'm trying to solve. How can I fix my little function to plot the derivative? Or is there a built-in function to do it? All of the built-ins (such as scipy.misc.derivative 取决于了解您要区分的函数。

Answer 1

在 Python v3.x range() 中创建一个 range object instead of a list. Apparently range objects don't support arithmetic. Try using a numpy ndarray。

x = np.arange(nodisk_data.shape[1])

numpy 数值微分

numpy numerical differentiation

python

numerical

numpy

numerical-methods