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
|
#include <boost/test/unit_test.hpp>
#include <aocommon/threadpool.h>
#include "../../math/tophatconvolution.h"
using aocommon::Image;
namespace wsclean {
namespace {
constexpr float v = 1.0f / 29.0f;
const Image reference_7x7_radius_3(
7, 7,
{
0.0f, 0.0f, 0.0f, v, 0.0f, 0.0f, 0.0f, //
0.0f, v, v, v, v, v, 0.0f, //
0.0f, v, v, v, v, v, 0.0f, //
v, v, v, v, v, v, v, //
0.0f, v, v, v, v, v, 0.0f, //
0.0f, v, v, v, v, v, 0.0f, //
0.0f, 0.0f, 0.0f, v, 0.0f, 0.0f, 0.0f //
});
float ref_value(ssize_t x, ssize_t y) {
const ssize_t width = reference_7x7_radius_3.Width();
const ssize_t height = reference_7x7_radius_3.Height();
if (x < 0 || y < 0 || x >= width || y >= height)
return 0.0;
else
return reference_7x7_radius_3[x + y * width];
}
} // namespace
BOOST_AUTO_TEST_SUITE(tophat_convolution)
BOOST_AUTO_TEST_CASE(make_tophat_with_large_radius) {
constexpr size_t kWidth = 10;
constexpr size_t kHeight = 12;
constexpr double kRadius = 15.0;
const Image image =
tophat_convolution::MakeTopHatImage(kWidth, kHeight, kRadius);
for (float value : image) {
BOOST_CHECK_CLOSE_FRACTION(value, 1.0 / (kWidth * kHeight), 1e-6);
}
}
BOOST_AUTO_TEST_CASE(make_single_pixel_tophat) {
constexpr size_t kWidth = 3;
constexpr size_t kHeight = 3;
constexpr double kRadius = 0.0;
const Image image =
tophat_convolution::MakeTopHatImage(kWidth, kHeight, kRadius);
for (size_t i = 0; i != kWidth * kHeight; ++i) {
const float expected =
(i == kWidth * ((kHeight - 1) / 2) + ((kWidth - 1) / 2)) ? 1.0f : 0.0f;
BOOST_CHECK_CLOSE_FRACTION(image[i], expected, 1e-6);
}
}
BOOST_AUTO_TEST_CASE(make_medium_tophat) {
constexpr size_t kWidth = 7;
constexpr size_t kHeight = 7;
// Use value slightly larger than 3 to avoid rounding errors for pixels that
// are exactly a distance of 3 pixels away.
constexpr double kRadius = 3.01;
const Image image =
tophat_convolution::MakeTopHatImage(kWidth, kHeight, kRadius);
for (size_t i = 0; i != kWidth * kHeight; ++i) {
BOOST_CHECK_CLOSE_FRACTION(image[i], reference_7x7_radius_3[i], 1e-6);
}
}
BOOST_AUTO_TEST_CASE(convolve_single_pixel) {
constexpr size_t kWidth = 7;
constexpr size_t kHeight = 7;
constexpr double kRadius = 3.01;
Image image(kWidth, kHeight, 0.0);
image[kWidth * ((kHeight - 1) / 2) + (kWidth - 1) / 2] = 3.0;
aocommon::ThreadPool::GetInstance().SetNThreads(2);
tophat_convolution::Convolve(image, kRadius);
for (size_t i = 0; i != kWidth * kHeight; ++i) {
// BOOST_CHECK_CLOSE_FRACTION doesn't work when comparing zeros to tiny
// non-zero values
BOOST_CHECK_LT(std::fabs(image[i] - 3.0 * reference_7x7_radius_3[i]), 1e-6);
}
}
BOOST_AUTO_TEST_CASE(convolve_multi_pixel) {
constexpr size_t kWidth = 35;
constexpr size_t kHeight = 35;
constexpr double kRadius = 3.01;
const ssize_t mid_x = (kWidth - 1) / 2;
const ssize_t mid_y = (kHeight - 1) / 2;
const ssize_t ref_mid_x = (reference_7x7_radius_3.Width() - 1) / 2;
const ssize_t ref_mid_y = (reference_7x7_radius_3.Height() - 1) / 2;
Image image(kWidth, kHeight, 0.0);
image[mid_y * kWidth + mid_x] = 3.0;
image[(mid_y + 1) * kWidth + mid_x + 1] = 1.0;
image[(mid_y + 2) * kWidth + mid_x + 1] = -2.0;
aocommon::ThreadPool::GetInstance().SetNThreads(2);
tophat_convolution::Convolve(image, kRadius);
for (ssize_t y = 0; y != kHeight; ++y) {
for (ssize_t x = 0; x != kWidth; ++x) {
const float value_a =
ref_value(x - mid_x + ref_mid_x, y - mid_y + ref_mid_y) * 3.0;
const float value_b =
ref_value(x - 1 - mid_x + ref_mid_x, y - 1 - mid_y + ref_mid_y) * 1.0;
const float value_c =
ref_value(x - 1 - mid_x + ref_mid_x, y - 2 - mid_y + ref_mid_y) *
-2.0;
const float reference = value_a + value_b + value_c;
BOOST_CHECK_LT(std::fabs(image[y * kWidth + x] - reference), 1e-6);
}
}
}
BOOST_AUTO_TEST_SUITE_END()
} // namespace wsclean
|
