Edge extraction based on Halcon learning [2] circles Hdev routine

The edges in the image are divided into lines and circles , For edges that are part of a circle , The circle parameters are estimated and the resulting circle is displayed .

* Read the picture 
read_image (Image, 'double_circle')

* close window 
dev_close_window ()
* Get the size of the picture 
get_image_size (Image, Width, Height)
* open windows 
dev_open_window (0, 0, Width, Height, 'black', WindowHandle)

* Split the area containing edges 
* Fast binarization 
fast_threshold (Image, Region, 0, 120, 7)
* Get the edge of the area --- Outline the area 
boundary (Region, RegionBorder, 'inner')
* Relative to its smallest surrounding rectangular clipping area .
clip_region_rel (RegionBorder, RegionClipped, 5, 5, 5, 5)
* inflation 
dilation_circle (RegionClipped, RegionDilation, 2.5)
* Cutout 
reduce_domain (Image, RegionDilation, ImageReduced)

* In the image subdomain containing the edge , Extract sub-pixel precise edge .

* Use Deriche、Lanser、Shen or Canny The filter extracts sub-pixel precise edges .
edges_sub_pix (ImageReduced, Edges, 'canny', 2, 20, 60)
* Sub pixel contour lines are divided into line segments and arc or elliptical arc .
segment_contours_xld (Edges, ContoursSplit, 'lines_circles', 5, 4, 3)
* Count 
count_obj (ContoursSplit, Number)
dev_display (Image)
dev_set_draw ('margin')
dev_set_color ('white')
dev_update_window ('off')
for I := 1 to Number by 1
    * Select the specified element in the array through the index number 
    select_obj (ContoursSplit, ObjectSelected, I)
    * return xld The global attribute value of the profile 
    get_contour_global_attrib_xld (ObjectSelected, 'cont_approx', Attrib)
    
    *Attrib Yes -1 0 1
    * Fit the circle to the arc segment of the circle 
    if (Attrib > 0)
        * Yes XLD Arc fitting of contour 
        fit_circle_contour_xld (ObjectSelected, 'ahuber', -1, 2, 0, 3, 2, Row, Column, Radius, StartPhi, EndPhi, PointOrder)
        * Draw the fitted circle 
        gen_circle_contour_xld (ContCircle, Row, Column, Radius, 0, rad(360), 'positive', 1.0)
        * Display circle 
        dev_display (ContCircle)
    endif
endfor
* Set the color 、 Line width 
dev_set_colored (12)
dev_set_line_width (3)
* Show edges  
dev_display (ContoursSplit)

---

* Read the picture 
read_image (Image, 'double_circle')

* close window 
dev_close_window ()

* Get the size of the picture 
get_image_size (Image, Width, Height)

* open windows 
dev_open_window (0, 0, Width, Height, 'black', WindowHandle)

* Split the area containing edges 
* Fast binarization 
fast_threshold (Image, Region, 0, 120, 7)

* Get the edge of the area --- Outline the area 
boundary (Region, RegionBorder, 'inner')

* inflation 
dilation_circle (RegionClipped, RegionDilation, 2.5)
* Cutout 
reduce_domain (Image, RegionDilation, ImageReduced)

* Use Deriche、Lanser、Shen or Canny The filter extracts sub-pixel precise edges .
edges_sub_pix (ImageReduced, Edges, 'canny', 2, 20, 60)

* Sub pixel contour lines are divided into line segments and arc or elliptical arc .
segment_contours_xld (Edges, ContoursSplit, 'lines_circles', 5, 4, 3)

* Count 
count_obj (ContoursSplit, Number)

dev_display (Image)
dev_set_draw ('margin')
dev_set_color ('white')
dev_update_window ('off')

for I := 1 to Number by 1
    * Select the specified element in the array through the index number 
    select_obj (ContoursSplit, ObjectSelected, I)

    * return xld The global attribute value of the profile 
    get_contour_global_attrib_xld (ObjectSelected, 'cont_approx', Attrib)
    
    *Attrib Yes -1 0 1
    * Fit the circle to the arc segment of the circle 
    if (Attrib > 0)
        * Yes XLD Arc fitting of contour 
        fit_circle_contour_xld (ObjectSelected, 'ahuber', -1, 2, 0, 3, 2, Row, Column, Radius, StartPhi, EndPhi, PointOrder)

        * Draw the fitted circle 
        gen_circle_contour_xld (ContCircle, Row, Column, Radius, 0, rad(360), 'positive', 1.0)
        * Display circle 
        dev_display (ContCircle)
    endif
endfor
* Set the color 、 Line width 
dev_set_colored (12)
dev_set_line_width (3)
* Show edges  
dev_display (ContoursSplit)

 

---

clip_region_rel(Region : RegionClipped : Top, Bottom, Left, Right : )

function ： Relative to its smallest surrounding rectangular clipping area .

Region ： Input area ;RegionClipped ： The output area is reduced ;Top, Bottom, Left, Right： Internal shrinkage .

Reduce the area by eliminating the part near the smallest surrounding rectangle of the area . say concretely , This means that the area is clipped small

A rectangle equal to or greater than the smallest surrounding rectangle

edges_sub_pix(Image : Edges : Filter, Alpha, Low, High : )

function ：  Use Deriche,Lanser,Shen or Canny The filter extracts sub-pixel precise edges

Image ： The input image ;Edges ： The edge of the output ;Alpha： Filter width ;

Low： Lower threshold for hysteresis threshold operation ;High： Upper threshold limit for hysteresis threshold operation

egment_contours_xld(Contours : ContoursSplit :Mode, SmoothCont, MaxLineDist1, MaxLineDist2 : )

function ： Division XLD The outline is a line segment and an arc or an elliptical arc

Contours : Enter the profile ;ContoursSplit： The contour after segmentation ;Mode： Contour segmentation mode ;

SmoothCont： The number of points used to smooth the contour ;MaxLineDist1： The maximum distance between the contour and the approximate line （ for the first time

iteration ）;MaxLineDist2 ： The maximum distance between the contour and the approximate line （ Second iteration ）