识别二进制图像中的曲线

Question

I have a few screenshots from experiments which display curves, and I want to extract that curve into an array to analyze the data.

The first part of my method is to transform the image into a binary one with a well chosen threshold that only leaves the curve. This works fine, but the second part consisting in extracting an array from the curve doesn't. I tried the "naive" method of going through the picture and identifying the white pixels, but the results are nowhere near what's expected.

Does someone have a solution?

import numpy as np
import matplotlib.pyplot as plt
import PIL.Image as img

gray = img.open('0100k.jpg').convert('L')
bw = gray.point(lambda x: 0 if x<240 else 255, '1')
bw = bw.convert('1')
bw.show()
pix = np.array(bw)
(n,m) = pix.shape
curve = np.zeros(n)
for i in range(n):
    for j in range(m):
        if pix[i,j] == 1:
        curve[i] = j
plt.plot(curve)
plt.show()

Answer 1

Thanks guys, the method now works, the solution was in the correct assignation of i and j to their respective dimensions. Thanks to @Matt Cremens for pointing it out.
But the plot thickens, it detects some non-existing pixels, like the ones visibles in the left side of the picture : http://s31.postimg.org/i0conbbxn/figure_1.png