Computer-Vision---Corner-Detector---Descriptor---Matcher/Matcher.py at master · duycao2506/Computer-Vision---Corner-Detector---Descriptor---Matcher · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
#Cao Khac Le Duy @copyright
# 1351008
# All rights are reserved


import sys
import cv2
import numpy as np
from enum import Enum
import time

class Matcher(object):
    def __init__(self):
        self.detector = None
        self.imgs = []
        pass

    def match(self,imgs):
        self.imgs = imgs
        pass

    def setDetector(self,detector):
        self.detector = detector
        detector.callback = self.doMatch
        pass

    def doMatch(self):
        pass

class SiftMatcher(Matcher):
    def __init__(self):
        super(SiftMatcher,self).__init__()


    def match(self,imgs):
        Matcher.match(self,imgs)
        self.doMatch()
        self.detector.addUi()
        pass

    def doMatch(self):
        kpts1 = self.detector.getDetected(self.imgs[0])
        kpts2 = self.detector.getDetected(self.imgs[1])
        grays = [cv2.cvtColor(img, cv2.COLOR_RGB2GRAY) for img in self.imgs]

        sift = cv2.xfeatures2d.SIFT_create()
        start = time.time()
        desc = []
        desc = sift.compute(grays,[kpts1, kpts2],desc)


        bf = cv2.BFMatcher()
        matches = None
        # print (desc[1][0][0])
        matches = bf.knnMatch(desc[1][0],desc[1][1], 2)

        # Apply ratio test
        good = []
        for m,n in matches:
            if m.distance < 0.75*n.distance:
                good.append([m])
        end = time.time()
        print end - start, " Time consuming "
        # cv2.drawMatchesKnn expects list of lists as matches.
        img3 = None
        img3 = cv2.drawMatchesKnn(self.imgs[0], kpts1, self.imgs[1], kpts2, matches,img3,matchColor = [0,255,0], flags=2)
        cv2.imshow(self.detector.windowName, img3)


class LbpMatcher(Matcher):
    def __init__(self):
        super(LbpMatcher, self).__init__()

    def match(self,imgs):
        Matcher.match(self,imgs)
        self.doMatch()
        self.detector.addUi()
        pass

    def doMatch(self):
        kpnts = [self.detector.getDetected(image) for image in self.imgs]
        grays = [cv2.cvtColor(img, cv2.COLOR_RGB2GRAY) for img in self.imgs]

        start = time.time()
        lbp = LbpDescriptorExtractor()
        desc = [lbp.extractDescriptors(kpnts[0],grays[0]),
                lbp.extractDescriptors(kpnts[1],grays[1])]
        bf = cv2.BFMatcher()

        matches = None
        matches = bf.knnMatch(np.asarray(desc[0],dtype=np.float32),np.asarray(desc[1],dtype=np.float32),2)

        # Apply ratio test
        good = []
        for m,n in matches:
            if m.distance < 0.75*n.distance:
                good.append([m])
        end = time.time()
        print end - start, " Time consuming "
        # cv2.drawMatchesKnn expects list of lists as matches.
        img3 = None
        img3 = cv2.drawMatchesKnn(self.imgs[0],kpnts[0],self.imgs[1],kpnts[1],matches,img3,matchColor = [0,255,0], flags=2)
        cv2.imshow(self.detector.windowName, img3)


class LbpDescriptorExtractor(object):
    "Extract descriptors from keypoints"

    def __init__(self):
        pass

    def extractDescriptors(self, kpnts, img):
        descrpt = [self.buildHistoVector(self.getCellLbp(key,img)) for key in kpnts]
        return descrpt

    def getCellLbp(self, keypoint, img):
        x,y = keypoint.pt
        gapx = len(img) - int(x)
        gapy = len(img[0]) - int(y)
        gapx =  gapx if gapx < 8 else 8
        gapy =  gapy if gapy < 8 else 8
        cell = img[int(x) - gapx : int(x) + gapx, int(y) - gapy : int(y) + gapy]
        return cell

    def buildHistoVector(self, lbpCell):
        vectorLbp = [0]*256
        for i in range(1,len(lbpCell)-1):
            for j in range(1,len(lbpCell[0])-1):
                index = self.getBinaryNumberAt(i,j,lbpCell)
                vectorLbp[index] += 1.0
        return vectorLbp

    def getBinaryNumberAt(self, x, y, ofCell):
        count = 0
        count += ((ofCell[x][y] < ofCell[x-1][y-1]) << 7)
        count += ((ofCell[x][y] < ofCell[x-1][y]) << 6)
        count += ((ofCell[x][y] < ofCell[x-1][y+1]) << 5)
        count += ((ofCell[x][y] < ofCell[x][y+1]) << 4)
        count += ((ofCell[x][y] < ofCell[x+1][y+1]) << 3)
        count += ((ofCell[x][y] < ofCell[x+1][y]) << 2)
        count += ((ofCell[x][y] < ofCell[x+1][y-1]) << 1)
        count += ((ofCell[x][y] < ofCell[x][y-1]) << 0)
        return count