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
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
/*
* video_homography.cpp
*
* Created on: Oct 18, 2010
* Author: erublee
*/
#include "opencv2/calib3d/calib3d.hpp"
#include "opencv2/highgui/highgui.hpp"
#include "opencv2/imgproc/imgproc.hpp"
#include "opencv2/features2d/features2d.hpp"
#include <iostream>
#include <list>
#include <vector>
using namespace std;
using namespace cv;
static void help(char **av)
{
cout << "\nThis program demonstrated the use of features2d with the Fast corner detector and brief descriptors\n"
<< "to track planar objects by computing their homography from the key (training) image to the query (test) image\n\n" << endl;
cout << "usage: " << av[0] << " <video device number>\n" << endl;
cout << "The following keys do stuff:" << endl;
cout << " t : grabs a reference frame to match against" << endl;
cout << " l : makes the reference frame new every frame" << endl;
cout << " q or escape: quit" << endl;
}
namespace
{
void drawMatchesRelative(const vector<KeyPoint>& train, const vector<KeyPoint>& query,
std::vector<cv::DMatch>& matches, Mat& img, const vector<unsigned char>& mask = vector<
unsigned char> ())
{
for (int i = 0; i < (int)matches.size(); i++)
{
if (mask.empty() || mask[i])
{
Point2f pt_new = query[matches[i].queryIdx].pt;
Point2f pt_old = train[matches[i].trainIdx].pt;
cv::line(img, pt_new, pt_old, Scalar(125, 255, 125), 1);
cv::circle(img, pt_new, 2, Scalar(255, 0, 125), 1);
}
}
}
//Takes a descriptor and turns it into an xy point
void keypoints2points(const vector<KeyPoint>& in, vector<Point2f>& out)
{
out.clear();
out.reserve(in.size());
for (size_t i = 0; i < in.size(); ++i)
{
out.push_back(in[i].pt);
}
}
//Takes an xy point and appends that to a keypoint structure
void points2keypoints(const vector<Point2f>& in, vector<KeyPoint>& out)
{
out.clear();
out.reserve(in.size());
for (size_t i = 0; i < in.size(); ++i)
{
out.push_back(KeyPoint(in[i], 1));
}
}
//Uses computed homography H to warp original input points to new planar position
void warpKeypoints(const Mat& H, const vector<KeyPoint>& in, vector<KeyPoint>& out)
{
vector<Point2f> pts;
keypoints2points(in, pts);
vector<Point2f> pts_w(pts.size());
Mat m_pts_w(pts_w);
perspectiveTransform(Mat(pts), m_pts_w, H);
points2keypoints(pts_w, out);
}
//Converts matching indices to xy points
void matches2points(const vector<KeyPoint>& train, const vector<KeyPoint>& query,
const std::vector<cv::DMatch>& matches, std::vector<cv::Point2f>& pts_train,
std::vector<Point2f>& pts_query)
{
pts_train.clear();
pts_query.clear();
pts_train.reserve(matches.size());
pts_query.reserve(matches.size());
size_t i = 0;
for (; i < matches.size(); i++)
{
const DMatch & dmatch = matches[i];
pts_query.push_back(query[dmatch.queryIdx].pt);
pts_train.push_back(train[dmatch.trainIdx].pt);
}
}
void resetH(Mat&H)
{
H = Mat::eye(3, 3, CV_32FC1);
}
}
int main(int ac, char ** av)
{
if (ac != 2)
{
help(av);
return 1;
}
BriefDescriptorExtractor brief(32);
VideoCapture capture;
capture.open(atoi(av[1]));
if (!capture.isOpened())
{
help(av);
cout << "capture device " << atoi(av[1]) << " failed to open!" << endl;
return 1;
}
cout << "following keys do stuff:" << endl;
cout << "t : grabs a reference frame to match against" << endl;
cout << "l : makes the reference frame new every frame" << endl;
cout << "q or escape: quit" << endl;
Mat frame;
vector<DMatch> matches;
BFMatcher desc_matcher(NORM_HAMMING);
vector<Point2f> train_pts, query_pts;
vector<KeyPoint> train_kpts, query_kpts;
vector<unsigned char> match_mask;
Mat gray;
bool ref_live = true;
Mat train_desc, query_desc;
const int DESIRED_FTRS = 500;
GridAdaptedFeatureDetector detector(new FastFeatureDetector(10, true), DESIRED_FTRS, 4, 4);
Mat H_prev = Mat::eye(3, 3, CV_32FC1);
for (;;)
{
capture >> frame;
if (frame.empty())
break;
cvtColor(frame, gray, COLOR_RGB2GRAY);
detector.detect(gray, query_kpts); //Find interest points
brief.compute(gray, query_kpts, query_desc); //Compute brief descriptors at each keypoint location
if (!train_kpts.empty())
{
vector<KeyPoint> test_kpts;
warpKeypoints(H_prev.inv(), query_kpts, test_kpts);
Mat mask = windowedMatchingMask(test_kpts, train_kpts, 25, 25);
desc_matcher.match(query_desc, train_desc, matches, mask);
drawKeypoints(frame, test_kpts, frame, Scalar(255, 0, 0), DrawMatchesFlags::DRAW_OVER_OUTIMG);
matches2points(train_kpts, query_kpts, matches, train_pts, query_pts);
if (matches.size() > 5)
{
Mat H = findHomography(train_pts, query_pts, RANSAC, 4, match_mask);
if (countNonZero(Mat(match_mask)) > 15)
{
H_prev = H;
}
else
resetH(H_prev);
drawMatchesRelative(train_kpts, query_kpts, matches, frame, match_mask);
}
else
resetH(H_prev);
}
else
{
H_prev = Mat::eye(3, 3, CV_32FC1);
Mat out;
drawKeypoints(gray, query_kpts, out);
frame = out;
}
imshow("frame", frame);
if (ref_live)
{
train_kpts = query_kpts;
query_desc.copyTo(train_desc);
}
char key = (char)waitKey(2);
switch (key)
{
case 'l':
ref_live = true;
resetH(H_prev);
break;
case 't':
ref_live = false;
train_kpts = query_kpts;
query_desc.copyTo(train_desc);
resetH(H_prev);
break;
case 27:
case 'q':
return 0;
break;
}
}
return 0;
}