halcon中variation_model_single实例注释.

 
* This example shows how to employ the new extensions of HALCON's variation model operators
* to perform customary print quality tests.
* In this example the variation model is built upon a single reference image.
* The example consists of three steps:
* 1. align the print objects similar to the reference image using a shape-based model
* 2. define the variation image by smoothing the object's contours
* 3. create the variation model
* Whether a print is labelled as OK or not, depends upon the size (area) of the difference to the reference image
* 

dev_close_window ()

*读取图片

read_image (Image, 'pen/pen-01')

*获取大小

get_image_size (Image, Width, Height)

*打开窗口

dev_open_window (0, 0, Width, Height, 'black', WindowHandle)

*关闭窗口更新

dev_update_off ()

*设置字体

set_display_font (WindowHandle, 12, 'courier', 'true', 'false')

*显示图片

dev_display (Image)

* 
* segment the logo and create a shape model for the alignment

*二值化图片

threshold (Image, Region, 125, 255)

*填充图片

fill_up (Region, RegionFillUp)

*计算两图片的不同,得到中间灰色有图像的部分.

difference (RegionFillUp, Region, RegionDifference)

*将区域进行凸包变换

shape_trans (RegionDifference, LogoArea, 'convex')

*肿胀区域,图片reduce区域处理完成

dilation_circle (LogoArea, LogoArea, 7)

*将图片重新剪切为有显示的区域.

reduce_domain (Image, LogoArea, ImageReduced)

*创建模板

create_shape_model (ImageReduced, 'auto', -rad(10), rad(20), 'auto', 'auto', 'use_polarity', [40,50], 40, ShapeModelID)

*计算中心点和面积

area_center (LogoArea, Area, ModelRow, ModelColumn)

 

* 
* define the variation image by smoothing the dilated regions obtained from the object's contours:
* Besides a binomial filter a neat trick is applied to get smoothly "polished" regions along the contours.
* In particular, the edges are enlarged and after their conversion into a dilated region the image

* is zoomed back to its original size using a weighting that smoothes the images further.

*亚像素分割图片

edges_sub_pix (ImageReduced, Edges, 'sobel_fast', 0.5, 10, 20)

*创建一个变换矩阵

hom_mat2d_identity (HomMat2DIdentity)

*矩阵x,y放大

hom_mat2d_scale (HomMat2DIdentity, 4, 4, 0, 0, HomMat2DScale)

*变换矩阵

affine_trans_contour_xld (Edges, ZoomedEdges, HomMat2DScale)

*产生空白图片

gen_image_const (VarImageBig, 'byte', 4*Width, 4*Height)

*计算XLD的对象

count_obj (ZoomedEdges, NEdges)

for i := 1 to NEdges by 1

   *选择对象

    select_obj (ZoomedEdges, ObjectSelected, i)

   *得到XLD的XY坐标点集

    get_contour_xld (ObjectSelected, RowEdge, ColEdge)

   *根据XLD坐标点集生成多边形

    gen_region_polygon (Region1, RowEdge, ColEdge)

  *肿胀区域,变得圆滑一些.

    dilation_circle (Region1, RegionDilation, 2.5)

   *将区域画到新的空图片上.

    paint_region (RegionDilation, VarImageBig, VarImageBig, 255, 'fill')

endfor

*压缩图片大小及正常大小

zoom_image_size (VarImageBig, VarImageSmall, Width, Height, 'weighted')

*binomial平滑图片

binomial_filter (VarImageSmall, VarImage, 3, 3)

*建立一个可变化比较模板

create_variation_model (Width, Height, 'byte', 'direct', VarModelID)

*将image图片转换为可变化比较模板VarModelID

prepare_direct_variation_model (Image, VarImage, VarModelID, 15, 4)

*显示这个模板

dev_display (VarImage)

disp_message (WindowHandle, 'Variation Image', 'window', -1, -1, 'black', 'true')
disp_continue_message (WindowHandle, 'black', 'true')
stop ()
* 

* print inspection

*以下开始比较图片了

for i := 1 to 30 by 1

    *读取图片

    read_image (Image, 'pen/pen-'+i$'02d')

    * locate the logo and align it to the reference image

* 通过模板查找目标目标位置坐标和角度

    find_shape_model (Image, ShapeModelID, -rad(10), rad(20), 0.5, 1, 0.5, 'least_squares', 0, 0.9, Row, Column, Angle, Score)

    if (|Score| # 0)

       *找到了目标,就刚性变换

        vector_angle_to_rigid (Row, Column, Angle, ModelRow, ModelColumn, 0, HomMat2D)

        affine_trans_image (Image, ImageAffinTrans, HomMat2D, 'constant', 'false')

       *剪切图片

        reduce_domain (ImageAffinTrans, LogoArea, ImageReduced1)

      * 开始比较

        compare_ext_variation_model (ImageReduced1, RegionDiff, VarModelID, 'absolute')

        *连通区域

        connection (RegionDiff, ConnectedRegions)

        *通过面积选择区域,忽略太小的点啊啥的.

        select_shape (ConnectedRegions, SelectedRegions, 'area', 'and', 10, 99999)

        *以下开始显示结果了 

      

        dev_display (ImageAffinTrans)

        *计算查找到的不同区域个数并显示ng or pass

        count_obj (SelectedRegions, NDefects)
        if (NDefects > 0)
            dev_set_color ('red')
            dev_set_draw ('margin')
            dev_set_line_width (2)
            dev_display (SelectedRegions)
            dev_set_color ('green')
            dev_set_line_width (1)
            dev_display (Edges)
            disp_message (WindowHandle, 'Image check not OK', 'window', -1, -1, 'red', 'false')
        else
            disp_message (WindowHandle, 'Image check OK', 'window', -1, -1, 'green', 'false')
        endif
    endif
    disp_continue_message (WindowHandle, 'black', 'true')
    stop ()
endfor
* 

* clean up

*最后记得要清理垃圾

clear_shape_model (ShapeModelID) 
clear_variation_model (VarModelID)