基于Halcon学习的基于形状模板匹配【六】find_mirror_dies.hdev例程

*许多模具通常蚀刻在晶圆上。在单曲之前模具可以由模具粘合机拾取并放置在引线上框架，它们必须从晶圆上切下来。镜子模具是用于确定切割过程的参数。

*在这里，镜子模具的位置是使用基于形状的匹配。

---

1、读取图片

dev_update_off ()
*关闭窗口
dev_close_window ()
*读取图片
read_image (Image, 'wafer/wafer_mirror_dies_01')

*打开自适应图片的窗口
dev_open_window_fit_image (Image, 0, 0, -1, -1, WindowHandle)
*设置显示字体
set_display_font (WindowHandle, 16, 'mono', 'true', 'false')
*设置线的宽度
dev_set_line_width (3)

*显示图片
dev_display (Image)
disp_message (WindowHandle, 'Determine the position of mirror dies on the wafer', 'window', 12, 12, 'black', 'true')
disp_continue_message (WindowHandle, 'black', 'true')
stop ()

 2、创建模板

*抠图
gen_rectangle1 (Rectangle, 362, 212, 414, 262)
reduce_domain (Image, Rectangle, ImageReduced)

*创建模板
create_shape_model (ImageReduced, 'auto', rad(0), rad(1), 'auto', 'auto', 'use_polarity', 'auto', 'auto', ModelID)
*得到轮廓
get_shape_model_contours (ModelContours, ModelID, 1)

 3、开始进行匹配

*阅读图片并进行处理

*开始进行for循环
for Index := 1 to 4 by 1
    read_image (Image, 'wafer/wafer_mirror_dies_' + Index$'02')

    *确定后视镜模具的位置，并计算找到位置所需的运行时间
    count_seconds (S1)
    *开始匹配
    find_shape_model (Image, ModelID, rad(0), rad(1), 0.5, 0, 0.0, 'least_squares', 2, 0.5, Row, Column, Angle, Score)
    count_seconds (S2)
    Runtime := (S2 - S1) * 1000
    
    *显示结果
    gen_cross_contour_xld (Cross, Row, Column, 6, rad(45))
    dev_display (Image)
    dev_display_shape_matching_results (ModelID, 'lime green', Row, Column, Angle, 1, 1, 0)
    dev_set_color ('orange')
    dev_display (Cross)
    disp_message (WindowHandle, |Score| + ' mirror dies located in ' + Runtime$'.1f' + ' ms', 'window', 12, 12, 'black', 'true')
    if (Index != 4)
        disp_continue_message (WindowHandle, 'black', 'true')
        stop ()
    endif
endfor

 

---

注：此处有个特别的算子，使得结果更加清晰可见

*为每个输入点生成一个十字形状的XLD轮廓。
gen_cross_contour_xld (Cross, Row, Column, 6, rad(45))