如何使用 numpy 以行和列形式将数据保存在.csv 文件中

Question

I am trying to read and image using OpenCV and after reading that image I have got some data which I have to save in a CSV file using numpy. Here is the program:-

import cv2 as cv
import numpy as np
import os

img1 = cv.imread('C:/Users/sbans/Pictures/bird.jpg')
dataA1 = os.path.basename('C:/Users/sbans/Pictures/bird.jpg')
height, width, channels = img1.shape
dataA2 = height
dataA3 = width
dataA4 = channels
a = int(height/2)
b = int(width/2)
px1  = img1[a,b]
dataA5 = px1[0]
dataA6 = px1[1]
dataA7 = px1[2]
a = np.array([dataA1, dataA2, dataA3, dataA4, dataA5, dataA6, dataA7])

img2 = cv.imread('C:/Users/sbans/Pictures/cat.jpg')
dataB1 = os.path.basename('C:/Users/sbans/Pictures/cat.jpg')
height, width, channels = img2.shape
dataB2 = height
dataB3 = width
dataB4 = channels
a = int(height/2)
b = int(width/2)
px2 = img2[a,b]
dataB5 = px2[0]
dataB6 = px2[1]
dataB7 = px2[2]
b = np.array([dataB1, dataB2, dataB3, dataB4, dataB5, dataB6, dataB7])
np.savetxt("stats.csv", np.stack((a,b)), delimiter=",", fmt='%s')

This error is coming:-

Traceback (most recent call last):

File "C:\Users\sbans\Documents\demo_opencv.py", line 32, in np.savetxt("stats.csv", np.stack((a,b)), delimiter=",", fmt='%s')

File "< array_function internals>", line 6, in stack

File "C:\Users\sbans\AppData\Local\Programs\Python\Python37\lib\site-packages\numpy\core\shape_base.py", line 425, in stack

raise ValueError('all input arrays must have the same shape') ValueError: all input arrays must have the same shape

Answer 1

You could simplify the code a bit by defining a function

def get_array(file):
    img = cv.imread(file)
    basename = os.path.basename(file)
    height, width, channels = img.shape
    h = int(height/2)
    w = int(width/2)
    px  = img[h,w]
    return np.array([basename, height, width, channels, px[0], px[1], px[2]]) 

Then savetxt can accept a list of same-sized 1D arrays

a = get_array('C:/Users/sbans/Pictures\bird.jpg')
b = get_array('C:/Users/sbans/Pictures\cat.jpg') 

np.savetxt("stats.csv", (a, b), delimiter=",", fmt='%s')

Answer 2

The default behavior of np.savetxt method is to replace the existing file with a new data.

If you want to sequentially write the data to a file then you need to have a reference for that file then use it for np.savetxt .

For your case:

f = open('stats.csv','w')

...
np.savetxt(f, np.row_stack(np.column_stack((dataA1, dataA2, dataA3, dataA4, dataA5, dataA6, dataA7))), delimiter=",", fmt='%s')


...
np.savetxt(f, np.row_stack(np.column_stack((dataB1, dataB2, dataB3, dataB4, dataB5, dataB6, dataB7))), delimiter=",", fmt='%s')


f.close()