-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathcensus.cpp
151 lines (149 loc) · 5.08 KB
/
census.cpp
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
#include "census.h"
//kkk
census::census(int id) {
//cout << "creat_census_id: "<<id << endl;
}
census::~census() {
//cout << "delete census_obj" << endl;
}
void census::census_transform(cv::Mat input_image, cv::Mat &modified_image, int window_sizex, int window_sizey) {
int image_height = input_image.rows;
int image_width = input_image.cols;
modified_image = cv::Mat::zeros(image_height, image_width, CV_64F);
//-----------census变换 ---------------------------------
int offsetx = (window_sizex - 1) / 2;
int offsety = (window_sizey - 1) / 2;
for (int j = 0; j < image_width - window_sizex; j++)
{
for (int i = 0; i < image_height - window_sizey; i++)
{
unsigned long census = 0;
uchar current_pixel = input_image.at<uchar>(i + offsety, j + offsetx); //窗口中心像素
cv::Rect roi(j, i, window_sizex, window_sizey); //方形窗口
cv::Mat window(input_image, roi);
for (int a = 0; a < window_sizey; a++)
{
for (int b = 0; b < window_sizex; b++)
{
if (!(a == offsety && b == offsetx))//中心像素不做判断
{
census = census << 1; //左移1位
}
uchar temp_value = window.at<uchar>(a, b);
if (temp_value <= current_pixel) //当前像素小于中心像素 01
{
census += 1;
}
}
}
modified_image.at<double>(i + offsety, j + offsetx) = census;
}
}
}
unsigned char census::census_hanming_dist(long long PL, long long PR) {
// Fast Hamming distance algorithm
// number 大于0 代表不同两个census序列互相转换的次数
unsigned char number = 0;
long long v;
v = PL ^ PR; /* ^ 异或运算 不同为1 相同为0*/
while (v)
{
v &= (v - 1); /* & 与运算*/
number++;
}
return number;
}
cost_sift::cost_sift(int id) {
// constructor
}
cost_sift::~cost_sift() {
//delete the class's variable member
}
void cost_sift::sift_transform(const cv::Mat img0,const cv::Mat img1,
vector<vector<vector<float>>>& Desc0,vector<vector<vector<float>>>& Desc1)
{
cv::cuda::GpuMat left_img;
cv::cuda::GpuMat right_img;
cv::Mat img0_gray, img1_gray;
cvtColor(img0, img0_gray, CV_BGR2GRAY);
cvtColor(img1, img1_gray, CV_BGR2GRAY);
left_img.upload(img0_gray);
right_img.upload(img1_gray);
//下面这个函数的原型是：
//explicit SURF_CUDA(
//double _hessianThreshold, //SURF海森特征点阈值 越小检测的特征点越多
//int _nOctaves=4, //尺度金字塔个数
//int _nOctaveLayers=2, //每一个尺度金字塔层数
//bool _extended=false, //如果true那么得到的描述子是128维，否则是64维
//float _keypointsRatio=0.01f,
//bool _upright = false
//);
//要理解这几个参数涉及SURF的原理
int dim = 64;
cv::cuda::SURF_CUDA surf(50, 4, 3);
/*分配下面几个GpuMat存储keypoint和相应的descriptor*/
vector<cv::KeyPoint> key0(img0.rows*img0.cols);
vector<cv::KeyPoint> key1(img0.rows*img0.cols);
vector<cv::KeyPoint> key_demo;
cv::cuda::GpuMat keypt0, keypt1;
cv::cuda::GpuMat desc0, desc1;
int point = 0;
// 自定义特征点 即每个像素点 平行换行遍历方式 后续描述子的排列以此来决定
for (int i = 0; i < img0.rows; i++)
for (int j = 0; j < img0.cols; j++) {
key0[point].pt.x = j;
key0[point].pt.y = i;
key0[point].size = 40;
key0[point].octave = 0;
key1[point].pt.x = j;
key1[point].pt.y = i;
key1[point].size = 40;
key1[point].octave = 0;
point++;
}
surf.uploadKeypoints(key0, keypt0);
surf.uploadKeypoints(key1, keypt1);
//keypt0.upload(key0);
//keypt1.upload(key1);
//描述子生成 gpu_mat (i,j)=(w,h)与普通mat相反
vector<float> descriptor0,descriptor1;
surf(left_img, cv::cuda::GpuMat(), keypt0, desc0,true); //使用自定义的keypt
surf(right_img, cv::cuda::GpuMat(), keypt1, desc1,true);
surf.downloadDescriptors(desc0, descriptor0); // 下载描述子给vector
surf.downloadDescriptors(desc1, descriptor1);
// cout << keypt0.size() << endl;
// 将64维向量按照w*h的分布储存
int index = 0;
int counter0=0;
int counter1=0;
vector<vector<vector<float>>> desc0_wh; // w*h分布的点的向量
vector<vector<vector<float>>> desc1_wh; // w*h分布的点的向量
for (int i = 0; i < img0_gray.rows; i++) {
vector<vector<float>> desc0_point_row; // 每一行的每一个点的向量
vector<vector<float>> desc1_point_row;
for (int j = 0; j < img0_gray.cols; j++) {
vector<float> desc0_point; // 每64个为一个点的向量
vector<float> desc1_point;
for (int m = 0; m <dim ; m++) {
desc0_point.push_back(descriptor0[index]);
desc1_point.push_back(descriptor0[index]);
if(isnan(descriptor0[index]))
counter0++;
if(isnan(descriptor1[index]))
counter1++;
index++;
}
desc0_point_row.push_back(desc0_point);
desc1_point_row.push_back(desc1_point);
}
desc0_wh.push_back(desc0_point_row);
desc1_wh.push_back(desc1_point_row);
}
cout<<"counter0: "<<counter0<<endl;
cout<<"counter1: "<<counter1<<endl;
Desc0=desc0_wh;
Desc1=desc1_wh;
//cout << keypt0.size().width << endl;
//cout << desc0.size().width << endl;
//cout << "descriptor generate" << endl;
}