convert experimental images to gray scale values

source/video_proc.py

@@ -0,0 +1,131 @@
+# -*- coding: utf-8 -*-
+"""
+Created on Mon Dec 13 21:54:24 2021
+
+@author: Weibo
+"""
+
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""
+Created on Mon Oct 11 15:39:40 2019
+    load map and visualize 
+
+@author: Weibo
+
+"""
+# %% Import necessary packages
+import os 
+import sys
+import cv2
+import numpy as np
+from PIL import Image
+import pandas as pd
+import matplotlib
+import matplotlib.pyplot as plt
+
+
+# path = 'I:/Hongyu/'
+path = '/media/weibo/Work/Wang/'
+cam = cv2.VideoCapture(path + 'w-1.avi')
+
+try:      
+    # creating a folder named data
+    if not os.path.exists(path + 'frame_orig'):
+        os.makedirs(path + 'frame_orig')
+# if not created then raise error
+except OSError:
+    print ('Error: Creating directory of data')
+pathFrame = path + 'frame_orig/'
+cframe = 0
+while(cframe < 4000):
+    ret, frame = cam.read()
+    if ret:
+        name = pathFrame + 'frame' + '{0:04d}'.format(cframe) + '.jpg'
+        cv2.imwrite(name, frame)
+        cframe += 1
+    else:
+        break
+
+cam.release()
+cv2.destroyAllWindows()
+
+# %% convert image to data
+# import an image from a file
+img = Image.open(pathFrame + 'frame0000.jpg')
+# extracting pixel map
+imgarr = img.load()
+# width and height of the image
+width0, height0 = img.size
+grayarr = np.ones((width0, height0)) * 255
+for i in range(width0):
+    for j in range(height0):
+        # get the RGB pixel value
+        rc, gc, bc = img.getpixel((i, j))
+        # convert color to grayscale using formula
+        grayscale = int(0.299 * rc + 0.587 * gc + 0.114 * bc)
+        grayarr[i, height0 - 1 - j] = grayscale 
+# imgdata = np.asarray(imgarr, dtype='int32')
+grayarr = np.transpose(grayarr)
+grayfil = grayarr[0, 0]
+print("the cutoff value is ", grayfil)
+ind = np.argwhere(grayarr == grayfil)
+# graynew = grayarr[58:152 , :]
+newrg = 513
+graynew = grayarr[20:240, 401:800] / 255 * newrg
+height = np.shape(graynew)[0]
+width = np.shape(graynew)[1]
+# %% plot contour to examine the conversion
+wid = width # width  # 100
+hei = height  #  40
+x1 = np.linspace(0, wid - 1, width)
+y1 = np.linspace(0, hei - 1, height)
+x2, y2 = np.meshgrid(x1, y1)
+graynew[:50, 350:] = 0.0
+textsize = 12
+numsize = 10
+fig, ax = plt.subplots(figsize=(6.4, 2.8))
+matplotlib.rc("font", size=textsize)
+rg1 = np.linspace(0, newrg, 64)
+cbar = ax.contourf(x2, y2, graynew, cmap="rainbow", levels=rg1, extend='both')  # rainbow_r
+# ax.set_xlim(-10.0, 30.0)
+# ax.set_ylim(-3.0, 10.0)
+ax.tick_params(labelsize=numsize)
+ax.set_xlabel(r"$x$", fontsize=textsize)
+ax.set_ylabel(r"$y$", fontsize=textsize)
+# plt.gca().set_aspect("equal", adjustable="box")
+# Add colorbar
+rg2 = np.linspace(0, newrg, 4)
+cbaxes = fig.add_axes([0.17, 0.68, 0.20, 0.07])  # x, y, width, height
+cbaxes.tick_params(labelsize=numsize)
+cbar = plt.colorbar(cbar, cax=cbaxes, extendrect='False',
+                    orientation="horizontal", ticks=rg2)
+cbar.set_label(
+    r"$\langle \rho \rangle/\rho_{\infty}$", rotation=0, fontsize=textsize
+)
+plt.savefig(path + "test.png", dpi=600, bbox_inches="tight")
+plt.show()
+# %% save dataframe
+dirs = sorted(os.listdir(pathFrame))
+for kk in range(np.size(dirs)):
+    img = Image.open(pathFrame + dirs[kk])
+    imgarr = img.load()
+    width, height = img.size
+    grayarr = np.ones((width, height)) * 255
+    if (width == width0) and (height == height0):
+        pass
+    else:
+        sys.exit("The shape of " + dirs[i] + "is not compatible!")
+    for i in range(width):
+        for j in range(height):
+            rc, gc, bc = img.getpixel((i, j))
+            grayscale = int(0.299 * rc + 0.587 * gc + 0.114 * bc)
+            grayarr[i, height - 1 - j] = grayscale 
+    grayarr = np.transpose(grayarr)
+    graynew = grayarr[20:240, 401:800] / 255 * newrg
+    graynew[:50, 350:] = 0.0
+    frame_val = np.vstack([x2.ravel(order='F'), y2.ravel(order='F')])
+    frame_val = np.vstack([frame_val, graynew.ravel(order='F')])
+    datafra = pd.DataFrame(frame_val.T, columns=['x', 'y', 'gray'])
+    datafra.to_hdf(path + 'framedata/' + '{0:04d}'.format(kk) + '.h5', 
+                   'w', format='fixed')