Numpy TypeError：列表索引必須是整數，而不是元組，如何更改代碼？

Question

def on_pushButtonTwo_clicked(self): 
    img = np.asarray(Image.open("testtwo.tif").convert # read image, convert it('L'))        
    img = 1 * (img < 127) 

    areasplit = np.split(img.ravel(), 24) # here we are splitting converted to 1D array 
    for i in areasplit: 
    area = (i == 0).sum() 
    print area 

    maxr = areasplit[argmax(sum(areasplit == 0, axis = 0)), :] # selecting subarrays with min and max white pixel numbers 
    minr = areasplit[argmin(sum(areasplit == 0, axis = 0)), :] 

    distance = cityblock(maxr, minr) # computing distance between subarrays 
    print distance 


    Result: type error: areasplit[argmax(sum(areasplit == 0, axis=0)), :] 
    TypeError: list indices must be integers, not tuple 

有什麼辦法可以選擇具有最小和最大白色像素數的子陣列嗎？ 如何正確更改代碼？謝謝

Answer 1

areasplit是ndarray的列表，而不是ndarray，所以它不能採取第二個索引。這應該工作：

maxr = areasplit[argmax(sum(areasplit == 0, axis = 0))] 
minr = areasplit[argmin(sum(areasplit == 0, axis = 0))] 

或者，您ndarrays的列表轉換成2D ndarray這些調用之前：

areasplit = numpy.array(areasplit)