forked from kuronekodaisuki/StitchCam
-
Notifications
You must be signed in to change notification settings - Fork 0
/
MyBlender.cpp
150 lines (133 loc) · 3.02 KB
/
MyBlender.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
//
//
//
//#include <nmmintrin.h>
#include "MyBlender.h"
namespace cv {
namespace detail {
MyBlender::MyBlender(int try_gpu)
{
use_gpu = false;
if (try_gpu)
{
if (gpu::getCudaEnabledDeviceCount() > 0)
{
use_gpu = true;
}
}
}
void MyBlender::prepare(Rect dst_roi)
{
/*
dst_.create(dst_roi.size(), CV_8UC3);
dst_.setTo(Scalar::all(0));
if (use_gpu)
{
gpuDst_.upload(dst_);
}
*/
dst_mask_.create(dst_roi.size(), CV_8U);
dst_mask_.setTo(Scalar::all(0));
dst_roi_ = dst_roi;
}
void MyBlender::feed(const Mat &img, const Mat &mask, Point tl)
{
CV_Assert(img.type() == CV_8UC3 || img.type() == CV_16SC3); // 両方受け付ける
CV_Assert(mask.type() == CV_8U);
// ここで入力に応じてdst_を作成する
if (dst_.empty()) {
dst_.create(dst_roi_.size(), img.type());
dst_.setTo(Scalar::all(0));
//#ifndef JETSON_TK1
if (use_gpu)
{
gpuDst_.upload(dst_);
}
//#endif
}
int dx = tl.x - dst_roi_.x;
int dy = tl.y - dst_roi_.y;
if (use_gpu)
{
//#ifdef JETSON_TK1
// cv::gpu::device::cudaFeed(img, mask, dst_, dx, dy);
//#else
cv::gpu::device::cudaFeed(img, mask, gpuDst_, dx, dy);
//#endif
}
else if (img.type() == CV_8UC3)
{
#ifdef _OPENMP
# pragma omp parallel for
#endif
for (int y = 0; y < img.rows; ++y)
{
const Point3_<uchar> *src_row = img.ptr<Point3_<uchar> >(y);
const uchar *mask_row = mask.ptr<uchar>(y);
Point3_<uchar> *dst_row = dst_.ptr<Point3_<uchar> >(dy + y);
uchar *dst_mask_row = dst_mask_.ptr<uchar>(dy + y);
for (int x = 0; x < img.cols; ++x)
{
if (mask_row[x])
dst_row[dx + x] = src_row[x];
dst_mask_row[dx + x] |= mask_row[x];
}
}
}
else // CV_16SC3
{
#ifdef _OPENMP
# pragma omp parallel for
#endif
for (int y = 0; y < img.rows; ++y)
{
const Point3_<short> *src_row = img.ptr<Point3_<short> >(y);
const uchar *mask_row = mask.ptr<uchar>(y);
Point3_<short> *dst_row = dst_.ptr<Point3_<short> >(dy + y);
uchar *dst_mask_row = dst_mask_.ptr<uchar>(dy + y);
for (int x = 0; x < img.cols; ++x)
{
if (mask_row[x])
dst_row[dx + x] = src_row[x];
dst_mask_row[dx + x] |= mask_row[x];
}
}
}
}
void MyBlender::feed(const gpu::GpuMat &img, const gpu::GpuMat &mask, Point tl)
{
CV_Assert(img.type() == CV_8UC3);
CV_Assert(mask.type() == CV_8U);
// ここで入力に応じてdst_を作成する
if (dst_.empty()) {
dst_.create(dst_roi_.size(), img.type());
dst_.setTo(Scalar::all(0));
if (use_gpu)
{
gpuDst_.upload(dst_);
}
}
int dx = tl.x - dst_roi_.x;
int dy = tl.y - dst_roi_.y;
cv::gpu::device::cudaFeed(img, mask, gpuDst_, dx, dy);
}
void MyBlender::blend(Mat &dst, Mat &dst_mask)
{
if (use_gpu)
{
//#ifndef JETSON_TK1
gpuDst_.download(dst);
gpuDst_.release();
//#endif
}
else
{
dst_.setTo(Scalar::all(0), dst_mask_ == 0);
dst = dst_;
}
dst_mask = dst_mask_;
dst_.release();
dst_mask_.release();
}
} // namespace detail
} // namespace cv