有没有一种从多个帧中提取单个像素的有效方法？

这是一些读取视频的代码，从每一帧中获取 ROI，并将其存储在一个大数组中。

import numpy as np
import cv2 as cv

cap = cv.VideoCapture("somevideo.mkv")
assert cap.isOpened(),"video can't be opened... is the path correct?"

nframes = int(cap.get(cv.CAP_PROP_FRAME_COUNT))

assert nframes == 2500,f"you said 2500 frames but it's {nframes} frames"

roi_x,roi_y = 12,34 # top left
roi_width,roi_height = 400,150

cube = np.empty((nframes,roi_height,roi_width,3),dtype=np.uint8)
# 3-chAnnel: assuming it's RGB/BGR data,not gray

for i in range(nframes):
    rv,frame = cap.read()
    assert rv,"video ended before it said it would!"
    cube[i] = frame[roi_y:roi_y+roi_height,roi_x:roi_x+roi_width]

cap.release()

谢谢大家的回复。我今天对此进行了研究，并认为我对我的解决方案感到满意。我显然没有导入视频文件，而是从笔记本电脑上的文件夹中取出保存的帧，然后在时域中提取单个像素。虽然，接下来我将实现您的视频代码 @Christopher 并清理我的。谢谢！

import numpy as np
import cv2 
import glob


 ### Reading in all of the chessboard files ###
 cv_img = []

 for img in glob.glob("/Python/OpenCv/Images/chessboard/*.jpg"):
        n = cv.imread(img)
        cv_img.append(n)

   list1 = np.empty((len(cv_img),100,100),dtype=np.uint8)
   i = 0 

  for img in cv_img: 
 
      img = img[100:200,100:200]
      new_img = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
      list1[i] = new_img
      i = i+1
      img = img + 1 
 
      column = 0;
      pixels_clear = np.ones((1,13)) #no_of_frames,one column 5,1
      pixels = np.ones((1,13))

 for rows in range(0,100):

       column=0

      while column < 100: 
    
         
           n_frame = 0 
    
        
        for n_frame in range(0,13):
             
             pixels_clear[0,n_frame] =list1[n_frame,rows,column]
             n_frames = n_frames + 1
             
             column = column + 1
             pixels = np.vstack([pixels,pixels_clear])

"""