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
// 几何变换（纯 MoonBit）：仿射变换 + 任意角度旋转
// 使用双线性插值，边界 clamp 到边缘像素
///|
/// 双线性插值采样。在 (x, y) 处采样通道 c 的值。
fn sample_bilinear(
data : Array[Int],
w : Int,
h : Int,
ch : Int,
c : Int,
x : Float,
y : Float,
) -> Int {
let x0 = x.to_int()
let y0 = y.to_int()
let w_minus_1 = w - 1
let h_minus_1 = h - 1
let x0c = x0.clamp(min=0, max=w_minus_1)
let y0c = y0.clamp(min=0, max=h_minus_1)
let x1 = (x0 + 1).clamp(min=0, max=w_minus_1)
let y1 = (y0 + 1).clamp(min=0, max=h_minus_1)
let fx = x - Float::from_int(x0)
let fy = y - Float::from_int(y0)
let p00 = data[(y0c * w + x0c) * ch + c]
let p10 = data[(y0c * w + x1) * ch + c]
let p01 = data[(y1 * w + x0c) * ch + c]
let p11 = data[(y1 * w + x1) * ch + c]
let top = Float::from_int(p00) * (1.0F - fx) + Float::from_int(p10) * fx
let bot = Float::from_int(p01) * (1.0F - fx) + Float::from_int(p11) * fx
(top * (1.0F - fy) + bot * fy + 0.5F).to_int()
}
///|
/// 仿射变换。matrix = (a, b, c, d, e, f) 映射 dst→src：
/// src_x = a * dst_x + b * dst_y + c
/// src_y = d * dst_x + e * dst_y + f
/// 使用双线性插值，超出边界的像素 clamp 到边缘。
/// 输出尺寸为 new_w × new_h，通道数与输入相同。
pub fn warp_affine(
img : @types.Image,
matrix : (Float, Float, Float, Float, Float, Float),
new_w : Int,
new_h : Int,
) -> @types.Image {
let (a, b, c, d, e, f) = matrix
let ch = img.channels
let w = img.width
let h = img.height
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 dy in 0..<new_h {
let fdy = Float::from_int(dy)
for dx in 0..<new_w {
let sx = a * Float::from_int(dx) + b * fdy + c
let sy = d * Float::from_int(dx) + e * fdy + f
for ci in 0..<ch {
let val = sample_bilinear(data_int, w, h, ch, ci, sx, sy)
out[(dy * new_w + dx) * ch + ci] = val.clamp(min=0, max=255).to_byte()
}
}
}
@types.Image::{
width: new_w,
height: new_h,
channels: ch,
data: Bytes::from_array(out),
}
}
///|
/// 任意角度旋转（绕中心，保留原尺寸，角落填充黑色）。
/// angle 单位为度，正值逆时针旋转。使用双线性插值。
pub fn rotate(img : @types.Image, angle_degrees : Float) -> @types.Image {
let w = img.width
let h = img.height
let ch = img.channels
let rad = angle_degrees * 3.14159265F / 180.0F
let cos_a = @math.cosf(rad)
let sin_a = @math.sinf(rad)
let cx = Float::from_int(w - 1) / 2.0F
let cy = Float::from_int(h - 1) / 2.0F
// 逆变换：dst → src（旋转 -angle）
// src_x = cos*(dx-cx) + sin*(dy-cy) + cx
// src_y = -sin*(dx-cx) + cos*(dy-cy) + cy
// 即 matrix = (cos, sin, cx-cos*cx-sin*cy, -sin, cos, cy+sin*cx-cos*cy)
let tx = cx - cos_a * cx - sin_a * cy
let ty = cy + sin_a * cx - cos_a * cy
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 w_max = Float::from_int(w) - 0.5F
let h_max = Float::from_int(h) - 0.5F
let w_minus_1_f = Float::from_int(w - 1)
let h_minus_1_f = Float::from_int(h - 1)
let out : Array[Byte] = Array::make(w * h * ch, b'\x00')
for dy in 0..<h {
let fdy = Float::from_int(dy)
for dx in 0..<w {
let fdx = Float::from_int(dx)
let sx = cos_a * fdx + sin_a * fdy + tx
let sy = -sin_a * fdx + cos_a * fdy + ty
// 超出边界的像素填黑（用 0.5 容差避免浮点误差）
if sx < -0.5F || sx > w_max || sy < -0.5F || sy > h_max {
// 已经是 0
continue
}
let sx_c = sx.clamp(min=0.0F, max=w_minus_1_f)
let sy_c = sy.clamp(min=0.0F, max=h_minus_1_f)
for ci in 0..<ch {
let val = sample_bilinear(data_int, w, h, ch, ci, sx_c, sy_c)
out[(dy * w + dx) * ch + ci] = val.clamp(min=0, max=255).to_byte()
}
}
}
@types.Image::{
width: w,
height: h,
channels: ch,
data: Bytes::from_array(out),
}
}
///|
/// 2×3 仿射变换矩阵 (a, b, c, d, e, f)
/// x' = a*x + b*y + c
/// y' = d*x + e*y + f
pub(all) struct AffineMatrix {
a : Float
b : Float
c : Float
d : Float
e : Float
f : Float
}
///|
/// 从 3 对源点→目标点计算 2×3 仿射变换矩阵。
pub fn get_affine_transform(
src : Array[(Float, Float)],
dst : Array[(Float, Float)],
) -> AffineMatrix? {
if src.length() != 3 || dst.length() != 3 {
return None
}
// 6×6 线性方程组
let a : Array[Array[Float]] = Array::make(6, [])
let b : Array[Float] = Array::make(6, 0.0F)
for i in 0..<3 {
let (sx, sy) = src[i]
let (dx, dy) = dst[i]
a[i * 2] = [sx, sy, 1.0F, 0.0F, 0.0F, 0.0F]
b[i * 2] = dx
a[i * 2 + 1] = [0.0F, 0.0F, 0.0F, sx, sy, 1.0F]
b[i * 2 + 1] = dy
}
let sol = solve_linear_system(a, b, 6)
match sol {
None => None
Some(v) =>
Some(AffineMatrix::{
a: v[0],
b: v[1],
c: v[2],
d: v[3],
e: v[4],
f: v[5],
})
}
}
///|
/// 从中心点、角度和缩放计算 2×3 旋转矩阵。
/// angle 为弧度，scale 为缩放因子。
pub fn get_rotation_matrix_2d(
center : (Float, Float),
angle : Float,
scale : Float,
) -> AffineMatrix {
let (cx, cy) = center
let alpha = scale * @math.cosf(angle)
let beta = scale * @math.sinf(angle)
AffineMatrix::{
a: alpha,
b: beta,
c: (1.0F - alpha) * cx - beta * cy,
d: -beta,
e: alpha,
f: beta * cx + (1.0F - alpha) * cy,
}
}