Opencv read picture details

Recently, I was doing deep learning target detection , use opencv Image Reading , Do operations such as flipping , Before I found myself, I was opencv The understanding of data storage for reading pictures is not comprehensive , Now I would like to make a summary as follows ：
opencv The picture is read from the first pixel on the left , Because it is color RGB picture , So there are three values at each pixel ,(opencv Is in accordance with the BRG Read pixel values in the order of ), Go through each line in turn , Each line is a (width, 3) Of numpy Format of two-dimensional array , Finally, add the two-dimensional array of each row to a large array , Form a three array , And finally img.shape Output (height, width, 3) Such a three-dimensional array

take opencv The read picture becomes normal RGB Pattern ：

img_path = '000001.jpg'
img = cv2.imread(img_path)

# opencv Self contained  imshow  Function can be displayed normally  BGR  Format image 
cv2.imshow('res', img)
cv2.waitKey(0)

#  take  BGR  Turn into  RGB, use  plt  Show 
img_rgb = img[:, :, ::-1]	#  use numpy The array converts the last dimension 
img_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)	#  use cv The built-in conversion function 

plt.imshow(img_rgb)
plt.show()

Horizontal mirror flip

img_flip = cv2.flip(img, 1)
img_flip = img[:, ::-1, :]
img_combine = np.hstack((img, img_flip))

cv2.imshow('res', img_combine)
cv2.waitKey(0)

Vertical mirror flip

img_flip = cv2.flip(img, 0)
img_flip = img[::-1, :, :]
img_combine = np.hstack((img, img_flip))

cv2.imshow('res', img_combine)
cv2.waitKey(0)