dennis.py

# -*- coding: utf-8 -*-


import cv2
import numpy as np
from matplotlib import pyplot as plt
import os
from datetime import datetime
import random 

def isThisOkay(img):
    return bool(random.getrandbits(1))
    
def testMyImage(imgPath):

    print("Stuff")
    
    imgRaw = cv2.imread(imgPath, cv2.IMREAD_COLOR) #IMREAD_GRAYSCALE IMREAD_UNCHANGED IMREAD_COLOR
    imgRawOrig = imgRaw.copy()
    img_b,img_g,img_r = cv2.split(imgRaw)
    img_gray = cv2.cvtColor(imgRaw, cv2.COLOR_BGR2GRAY)
    
    equ = cv2.equalizeHist(img_r)
    
    #equ_th, dst = cv2.threshold(equ, 0, 200, cv2.THRESH_BINARY);
    
    th, bw = cv2.threshold(img_r, 0, 255, cv2.THRESH_BINARY | cv2.THRESH_OTSU)
    edges = cv2.Canny(imgRaw, th/2, th)
    
    plt.subplot(3,3,1),plt.imshow(imgRaw)
    plt.subplot(3,3,2),plt.imshow(bw, 'gray')
    
    # noise removal
    kernel = np.ones((13,13),np.uint8)
    opening = cv2.morphologyEx(bw, cv2.MORPH_OPEN,kernel, iterations = 2)
    # sure background area
    sure_bg = cv2.dilate(opening,kernel,iterations=3)
    # Finding sure foreground area
    dist_transform = cv2.distanceTransform(opening,cv2.DIST_L2,5)
    ret, sure_fg = cv2.threshold(dist_transform,0.4*dist_transform.max(),255,0)
    # Finding unknown region
    sure_fg = np.uint8(sure_fg)
    unknown = cv2.subtract(sure_bg,sure_fg)
    
    plt.subplot(3,3,4), plt.imshow(sure_bg)
    plt.subplot(3,3,5), plt.imshow(sure_fg)
    
    # Marker labelling
    num_labels, labels, stats, centroids = cv2.connectedComponentsWithStats(sure_fg)
    # Add one to all labels so that sure background is not 0, but 1
    labels = labels+1
    # Now, mark the region of unknown with zero
    labels[unknown==255] = 0
    
    
    labels = cv2.watershed(imgRaw, labels)
    imgRaw[labels == -1] = [255,0,0]
    
    plt.subplot(3,3,3),plt.imshow(labels)
    plt.subplot(3,3,6),plt.imshow(imgRaw)
    
    
    print("Number of markers: ", num_labels)
    
    newpath = "test/contures/" +  datetime.today().strftime("%d-%b-%Y_%Hh%Mm%S")
    if not os.path.exists(newpath):
        os.makedirs(newpath)
    
    for i in range(1, num_labels):
        if True: #i < 50:    
            newLabel = labels.copy()
            newLabel[newLabel != i] = 0
            newLabel[newLabel == i] = 255
            newLabel = newLabel.astype(np.uint8)
            th, imgloc = cv2.threshold(newLabel, 0, 255, cv2.THRESH_BINARY | cv2.THRESH_OTSU)
            image, contours, hierarchy = cv2.findContours(imgloc, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
        
            for cont in contours:
                x,y,w,h = cv2.boundingRect(cont)
                okayCheck = isThisOkay(imgRawOrig[y:y+h,x:x+w])
                if not okayCheck:
                    labels[labels == i] = 0
                    
                cv2.imwrite(newpath + '/' + str(i) +'.png', imgRawOrig[y:y+h,x:x+w])
    
    
    numNotBG = np.count_nonzero(bw.flatten())
    print("Not Background: ", numNotBG)
    labelCp = labels.copy()
    labelCp[labelCp < 2] = 0
    numOK = np.count_nonzero(labelCp.flatten())
    print("Good stuff: ", numOK)
    percOk = (numOK/numNotBG)*100.0
    print("OK: ", percOk)
    
    
    fred = np.fft.fft2(img_r)
    fshiftred = np.fft.fftshift(fred)
    magnitude_spectrum_red = 20*np.log(np.abs(fshiftred))
    plt.subplot(3,3,7),plt.imshow(magnitude_spectrum_red, 'gray')
    
    
    # Overlay image
    total = bw
    good = labels.copy()
    
    alpha = 0.5
    overlay = imgRawOrig.copy()
    output = imgRawOrig.copy()
    
    output[total > 0] = (0, 255, 0)
    output[labels > 1] = (0, 0, 255)
    
    cv2.addWeighted(overlay, alpha, output, 1 - alpha, 	0, output)
    
    
    plt.subplot(3,3,8),plt.imshow(output)
    
    
    plt.show()
    cv2.namedWindow("window", cv2.WINDOW_FREERATIO)
    cv2.imshow('window', output)
    cv2.imwrite('okay_reportSingle.jpg', output)
    print("WUHUUU")


if __name__ == "__main__":
    testMyImage('dorde/okay1.jpg')