Halcon 光度立体法应用（二）——皮革表面缺陷检测_halcon 3d检测凹坑凸起

Halcon 光度立体法应用——皮革表面缺陷检测

如果想深刻、系列的了解光度立体法，建议根据博客顺序观看。在这个例程中将会介绍通过光度立体法生成的图像适用场景。

总体代码注释说明

* 此例程介绍的是利用光度立体法对皮革样品表面进行缺陷检测。
* 
* 程序初始化
dev_update_off ()
dev_close_window ()
dev_open_window (0, 0, 640, 480, 'black', WindowHandle)
set_display_font (WindowHandle, 14, 'mono', 'true', 'false')
Message := 'Inspect leather samples using photometric stereo'
disp_message (WindowHandle, Message, 'window', 12, 12, 'black', 'true')
disp_continue_message (WindowHandle, 'black', 'true')
stop ()
* 
* 第一部分，利用反射率图像检测皮革表面缺陷
* 
* 展示不同方向光源成像图像
read_image (Images, 'photometric_stereo/leather_1_0' + [1:4])
for I := 1 to 4 by 1
    Message := 'Sample 1: Acquire image ' + I + ' of 4'
    select_obj (Images, ObjectSelected, I)
    dev_display (ObjectSelected)
    disp_message (WindowHandle, Message, 'window', 12, 12, 'black', 'true')
    wait_seconds (0.5)
endfor
* 
* 
* 应用光度立体法生成的反射率图进行缺陷检测
Tilts := [6.1,95.0,-176.1,-86.8]
Slants := [41.4,42.6,41.7,40.9]
ResultType := ['gradient','albedo']
photometric_stereo (Images, HeightField, Gradient, Albedo, Slants, Tilts, ResultType, 'poisson', [], [])
* 
* 显示反射率图
dev_display (Albedo)
disp_message (WindowHandle, 'The defect is clearly visible in the albedo image', 'window', 12, 12, 'black', 'true')
disp_continue_message (WindowHandle, 'black', 'true')
stop ()
* 
* 检测缺陷
* 
var_threshold (Albedo, Region, 15, 15, 0.4, 0.4, 'light')
connection (Region, ConnectedRegions)
select_shape (ConnectedRegions, SelectedRegions, 'area', 'and', 10, 99999)
union1 (SelectedRegions, RegionUnion)
closing_circle (RegionUnion, RegionClosing, 3.5)
connection (RegionClosing, Defects)
area_center (Defects, Area, Row, Column)
gen_circle (Circle, Row, Column, gen_tuple_const(|Row|,sqrt(Area) + 30))
* 缺陷显示
dev_display (Albedo)
dev_set_color ('red')
dev_set_draw ('margin')
dev_set_line_width (4)
dev_display (Circle)
disp_message (WindowHandle, 'Albedo image with defect', 'window', 12, 12, 'black', 'true')
disp_continue_message (WindowHandle, 'black', 'true')
stop ()
* 
* 第二部分区别第一部分缺陷类型，使用不同图像检测。
* 
* 展示不同方向光源成像图像
read_image (Images, 'photometric_stereo/leather_2_0' + [1:4])
for I := 1 to 4 by 1
    Message := 'Sample 2: Acquire image ' + I + ' of 4'
    select_obj (Images, ObjectSelected, I)
    dev_display (ObjectSelected)
    disp_message (WindowHandle, Message, 'window', 12, 12, 'black', 'true')
    wait_seconds (0.5)
endfor
* 
* 
* 根据光度立体法生成反射率图（反射率图无法检测）
photometric_stereo (Images, HeightField, Gradient, Albedo, Slants, Tilts, ResultType, 'poisson', [], [])
threshold (Albedo, Region1, 128, 255)
* 
* 显示反射率图
dev_display (Albedo)
Message := 'These defects are difficult to detect in the albedo image.'
Message[1] := 'Therefore, we use the gradient information to detect them.'
disp_message (WindowHandle, Message, 'window', 12, 12, 'black', 'true')
disp_continue_message (WindowHandle, 'black', 'true')
stop ()
* 
* 利用梯度信息检测纹理表面上小的突变区域缺陷。
* 
derivate_vector_field (Gradient, Curl, 1, 'curl')
derivate_gauss (Curl, CurlGradient, 1, 'gradient')
* 
* 以上对梯度信息图进行高斯导数卷积操作并显示处理后图像
dev_display (CurlGradient)
Message := 'Changes in the gradient curl'
disp_message (WindowHandle, Message, 'window', 12, 12, 'black', 'true')
disp_continue_message (WindowHandle, 'black', 'true')
stop ()
* 
* 检测纹理缺陷
threshold (CurlGradient, Region, 0, 0.01)
rank_region (Region, RegionCount, 10, 10, 30)
connection (RegionCount, ConnectedRegions)
select_shape (ConnectedRegions, SelectedRegions, 'area', 'and', 2000, 99999)
union1 (SelectedRegions, RegionUnion)
rank_region (RegionUnion, RegionCount1, 25, 25, 170)
connection (RegionCount1, NoTextured)
* 
* 显示缺陷
dev_display (Albedo)
dev_set_draw ('margin')
dev_set_color ('red')
dev_set_line_width (3)
dev_display (NoTextured)
disp_message (WindowHandle, 'Non-textured areas on leather', 'window', 12, 12, 'black', 'true')
stop ()
* 
* 利用曲率信息检测划痕，根据划痕表征的高曲率特性检测划痕。
* 
derivate_vector_field (Gradient, MeanCurvature, 1, 'mean_curvature')
* 
* 显示表面平均曲率图
dev_display (MeanCurvature)
Message := 'Mean curvature of the surface'
disp_message (WindowHandle, Message, 'window', 12, 12, 'black', 'true')
disp_message (WindowHandle, 'Press F5', 'image', 720, 850, 'black', 'true')
stop ()
* 
* 检测划痕
abs_image (MeanCurvature, ImageAbs)
threshold (ImageAbs, Region2, 0.15, 255)
connection (Region2, ConnectedRegions1)
select_shape (ConnectedRegions1, SelectedRegions1, 'area', 'and', 10, 99999)
union1 (SelectedRegions1, RegionUnion1)
closing_circle (RegionUnion1, RegionClosing, 1.5)
connection (RegionClosing, ConnectedRegions2)
select_shape (ConnectedRegions2, SelectedRegions2, 'max_diameter', 'and', 50, 99999)
select_gray (SelectedRegions2, MeanCurvature, SelectedRegions3, 'deviation', 'and', 0.2, 255)
* 
* 显示划痕
dev_display (Albedo)
dev_set_draw ('margin')
dev_set_color ('red')
dev_set_line_width (3)
dev_display (SelectedRegions3)
disp_message (WindowHandle, 'Deep scratch', 'window', 12, 12, 'black', 'true')
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

代码解析：

Part1利用反射率图检测皮革表面缺陷，而Part2却利用梯度信息图检测缺陷，Why？

仔细观察不难发现，Part1中的缺陷区域除了第一幅图（右边打光）之外其他展现的都是高亮特性。缺陷特征较背景区域具有较高的反光特性，所以反射率图能很好的凸显缺陷特征，所以用反射率图检测缺陷。
好奇的同志却在想，为啥不用梯度图呢？有想法是好的，那么假设我们用梯度信息图检测，效果会怎样？梯度信息图如下：
可以看出缺陷和纹理对比度很差，所以不能用梯度信息图检测Part1中的缺陷

Part2中的缺陷

从图中可以看出，缺陷表面的皮革纹理几乎不存在了，所以这种缺陷很可能是比较钝的物体造成。对于比较平缓的缺陷区域，梯度几乎没有变化，但对于背景纹理它们的梯度会始终存在，因此利用梯度信息图检测这种缺陷是一种不错的选择。划痕缺陷正好与其相反，划痕表征的是细长状，尖锐物体划伤，所以边缘梯度较锐利，同样用梯度信息提取。

总结：如果在实际应用中，有些产品缺陷对光源角度有要求，且方向不固定（比如：带方向的缺陷，需要多角度打光才能凸显缺陷的产品）那么就可以考虑光度立体法。