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
// 图像金字塔（纯 MoonBit）
///|
/// 高斯金字塔下采样。将图像尺寸减半（宽高各 /2），使用 2x2 均值滤波。
pub fn pyr_down(img : @types.Image) -> @types.Image raise @types.LoadError {
if img.channels != 1 && img.channels != 3 && img.channels != 4 {
raise @types.LoadError::DecodeFailed("pyr_down 仅支持 channels=1/3/4")
}
let ch = img.channels
let w = img.width
let h = img.height
let new_w = w / 2
let new_h = h / 2
if new_w == 0 || new_h == 0 {
raise @types.LoadError::DecodeFailed("图像太小，无法下采样")
}
let data_int : Array[Int] = Array::make(w * h * ch, 0)
for i in 0..<(w * h * ch) {
data_int[i] = img.data[i].to_int()
}
let out : Array[Byte] = Array::make(new_w * new_h * ch, b'\x00')
for y in 0..<new_h {
for x in 0..<new_w {
for c in 0..<ch {
let mut sum = 0
let mut count = 0
for dy in 0..<2 {
for dx in 0..<2 {
let sx = x * 2 + dx
let sy = y * 2 + dy
if sx < w && sy < h {
sum += data_int[(sy * w + sx) * ch + c]
count += 1
}
}
}
out[(y * new_w + x) * ch + c] = (sum / count).to_byte()
}
}
}
@types.Image::{
width: new_w,
height: new_h,
channels: ch,
data: Bytes::from_array(out),
}
}
///|
/// 拉普拉斯金字塔上采样。将图像尺寸加倍（宽高各 *2），使用双线性插值。
pub fn pyr_up(img : @types.Image) -> @types.Image raise @types.LoadError {
if img.channels != 1 && img.channels != 3 && img.channels != 4 {
raise @types.LoadError::DecodeFailed("pyr_up 仅支持 channels=1/3/4")
}
let ch = img.channels
let w = img.width
let h = img.height
let new_w = w * 2
let new_h = h * 2
let data_int : Array[Int] = Array::make(w * h * ch, 0)
for i in 0..<(w * h * ch) {
data_int[i] = img.data[i].to_int()
}
// 预计算 x 方向插值参数
let x0_arr : Array[Int] = Array::make(new_w, 0)
let x1_arr : Array[Int] = Array::make(new_w, 0)
let wx_arr : Array[Float] = Array::make(new_w, 0.0F)
for x in 0..<new_w {
let fx = (Float::from_int(x) - 0.5F) / 2.0F
let x0 = fx.to_int().clamp(min=0, max=w - 1)
x0_arr[x] = x0
x1_arr[x] = (x0 + 1).clamp(min=0, max=w - 1)
wx_arr[x] = fx - Float::from_int(x0)
}
// 预计算 y 方向插值参数
let y0_arr : Array[Int] = Array::make(new_h, 0)
let y1_arr : Array[Int] = Array::make(new_h, 0)
let wy_arr : Array[Float] = Array::make(new_h, 0.0F)
for y in 0..<new_h {
let fy = (Float::from_int(y) - 0.5F) / 2.0F
let y0 = fy.to_int().clamp(min=0, max=h - 1)
y0_arr[y] = y0
y1_arr[y] = (y0 + 1).clamp(min=0, max=h - 1)
wy_arr[y] = fy - Float::from_int(y0)
}
let out : Array[Byte] = Array::make(new_w * new_h * ch, b'\x00')
for y in 0..<new_h {
let y0 = y0_arr[y]
let y1 = y1_arr[y]
let wy = wy_arr[y]
let wy_inv = 1.0F - wy
let row0 = y0 * w
let row1 = y1 * w
for x in 0..<new_w {
let x0 = x0_arr[x]
let x1 = x1_arr[x]
let wx = wx_arr[x]
let wx_inv = 1.0F - wx
for c in 0..<ch {
let v00 = Float::from_int(data_int[(row0 + x0) * ch + c])
let v01 = Float::from_int(data_int[(row0 + x1) * ch + c])
let v10 = Float::from_int(data_int[(row1 + x0) * ch + c])
let v11 = Float::from_int(data_int[(row1 + x1) * ch + c])
let top = v00 * wx_inv + v01 * wx
let bot = v10 * wx_inv + v11 * wx
let v = top * wy_inv + bot * wy
let clamped = if v < 0.0F {
0
} else if v > 255.0F {
255
} else {
v.to_int()
}
out[(y * new_w + x) * ch + c] = clamped.to_byte()
}
}
}
@types.Image::{
width: new_w,
height: new_h,
channels: ch,
data: Bytes::from_array(out),
}
}
///|
/// 构建高斯金字塔。返回各级图像（从原图到最小级），levels 为最大层数（0=自动到 1x1）。
pub fn build_gaussian_pyramid(
img : @types.Image,
levels : Int,
) -> Array[@types.Image] raise @types.LoadError {
let result : Array[@types.Image] = [img]
let mut current = img
let max_levels = if levels == 0 { 100 } else { levels }
for _ in 1..<max_levels {
if current.width < 2 || current.height < 2 {
break
}
current = pyr_down(current)
result.push(current)
}
result
}
///|
/// 构建拉普拉斯金字塔。返回各级拉普拉斯图像（高斯金字塔相邻层之差）。
/// 最后一层为高斯金字塔的最小级（不加差）。
pub fn build_laplacian_pyramid(
img : @types.Image,
levels : Int,
) -> Array[@types.Image] raise @types.LoadError {
let gaussian = build_gaussian_pyramid(img, levels)
let result : Array[@types.Image] = []
for i in 0..<(gaussian.length() - 1) {
let upsampled = pyr_up(gaussian[i + 1])
let lap = laplacian_diff(gaussian[i], upsampled)
result.push(lap)
}
result.push(gaussian[gaussian.length() - 1])
result
}
///|
fn laplacian_diff(a : @types.Image, b : @types.Image) -> @types.Image {
let ch = a.channels
let w = a.width
let h = a.height
let a_int : Array[Int] = Array::make(w * h * ch, 0)
for i in 0..<(w * h * ch) {
a_int[i] = a.data[i].to_int()
}
let bw = b.width
let bh = b.height
let b_int : Array[Int] = Array::make(bw * bh * ch, 0)
for i in 0..<(bw * bh * ch) {
b_int[i] = b.data[i].to_int()
}
let out : Array[Byte] = Array::make(w * h * ch, b'\x00')
for y in 0..<h {
let by = y * bh / h
for x in 0..<w {
let bx = x * bw / w
for c in 0..<ch {
let va = a_int[(y * w + x) * ch + c]
let vb = b_int[(by * bw + bx) * ch + c]
let diff = va - vb + 128
let clamped = if diff < 0 { 0 } else if diff > 255 { 255 } else { diff }
out[(y * w + x) * ch + c] = clamped.to_byte()
}
}
}
@types.Image::{
width: w,
height: h,
channels: ch,
data: Bytes::from_array(out),
}
}