问照片相似性检查-将循环结果附加到数据帧

EN

data = {'Photo': ['C:\Lucy Maud in Garden.jpg','C:\Henry by car.jpg','C:\Lucy and Henry arms together.jpg','C:\Lucy Maud with dog.jpg'],
     'NearMatch': ['C:\Lucy Maud in Garden2.jpg','C:\Henry by car2.jpg','C:\Lucy and Henry arms together2.jpg','C:\Lucy Maud with dog2.jpg'],
     'Similarity': [1,2,1,11]
        }


df = pd.DataFrame (data, columns = ['Photo','NearMatch','Similarity'])

from __future__ import division

import cv2
import numpy as np
import glob
import pandas as pd

    # Sift and Flann
sift = cv2.SIFT_create()


index_params = dict(algorithm=0, trees=5)
search_params = dict()
flann = cv2.FlannBasedMatcher(index_params, search_params)

#prep the empty lists

countInner = 0
countOuter = 1
countNoMatch = 0
nearMatch = []
nearMatch2 = []
listOfSimilarities = []
listOfDisimilarities = []

# Load all the images

folder1 = r"C:/ProbablyDups/**"
folder2 = r"C:/DefinitiveCopy/**"


extensionsOnly = ('.jpeg','.jpg','.png','.tif','.tiff','.gif')

siftOut1 = {}
for a in glob.iglob(folder1,recursive=True):
    if not a.lower().endswith(extensionsOnly):
        continue
    image1 = cv2.imread(a)
    kp_1, desc_1 = sift.detectAndCompute(image1, None)
    siftOut1[a]=(kp_1,desc_1)

siftOut2 = {}
for a in glob.iglob(folder2,recursive=True):
    if not a.lower().endswith(extensionsOnly):
        continue
    image1 = cv2.imread(a)
    kp_1, desc_1 = sift.detectAndCompute(image1, None)
    siftOut2[a]=(kp_1,desc_1)

#Compare photos in loops
for a in glob.iglob(folder1,recursive=True):
    if not a.lower().endswith(extensionsOnly):
        continue

    (kp_1,desc_1) = siftOut1[a]

    for b in glob.iglob(folder2,recursive=True):


        if not b.lower().endswith(extensionsOnly):

            continue

        if b.lower().endswith(extensionsOnly):

            countInner += 1


        (kp_2,desc_2) = siftOut2[b]

        matches = flann.knnMatch(desc_1, desc_2, k=2)

        good_points = []

        for m, n in matches:
            if m.distance < 0.6*n.distance:
                good_points.append(m)

        number_keypoints = 0
        if len(kp_1) >= len(kp_2):
            number_keypoints = len(kp_1)
        else:
            number_keypoints = len(kp_2)

        percentage_similarity = int(float(len(good_points)) / number_keypoints * 100)
        # add a tick to the counter if there is positive match
        if percentage_similarity > 16:
            countNoMatch =+1
        #part that is not working:
        if percentage_similarity < 16 and percentage_similarity > 0:
            nearMatch.append(a)
            nearMatch2.append(b)
            listOfSimilarities.append(percentage_similarity)
    
    if countNoMatch == 0:
        listOfDisimilarities.append(a)
        df2=pd.DataFrame({"NoMatch":listOfDisimilarities})
        zippedList =  list(zip(nearMatch,nearMatch2, listOfSimilarities))
        print(zippedList)
        nearMatch = []
        nearMatch2 = []
        final_df = pd.concat(zippedList, ignore_index=True)
    
    countNoMatch = 0
    if a.lower().endswith(extensionsOnly):
        countOuter += 1
print(final_df)

df.to_csv(r"C:/Documents/NearMatch.csv")