Halcon practical: design idea of solder joint detection

One 、 summary

This article is about Blob Example one , The method of segmentation using gray threshold , The algorithm of solder joint detection .

Two 、 Problem analysis

The picture of the circuit board is as follows ： The solder joint is the black circular area above .

  Problem analysis ：

1） The area containing the solder joint is determined

The solder joint is a dark target area , The distribution area of solder joint is in the bright area W,, therefore , You need to put a large set M Give priority to .

* Grayscale processing   Two valued 
threshold (Bond, Bright, 100, 255)
* Morphological processing   Generally used to locate , Shape conversion ,rectangle2 Rectangle with direction 
shape_trans (Bright, Die, 'rectangle2')

2) Extract the background through shallow area M

3） adopt  shape_trans Take out the rectangle of the light colored area .

Be careful ： Then the area containing the foreground is brought forward , Narrowed the search ; Later, the operation will be carried out in a smaller area . adopt reduce_domain Take out the image in the area , Prepare for the second treatment .

reduce_domain (Bond, Die, DieGrey)

4） Extract the dark areas in the region

threshold (DieGrey, Wires, 0, 50)

5） Calculate the welding spot area , And get the area

* Fill the area with the given shape feature ： Area is 1~100 Region

fill_up_shape (Wires, WiresFilled, 'area', 1, 100)

6） Make the radius greater than 15.5 circular , Extract by open operation

opening_circle (WiresFilled, Balls, 15.5)

7） Separate the area

connection (Balls, SingleBalls)

8） Change roundness 0.85 The above areas are preserved

select_shape (SingleBalls, IntermediateBalls, 'circularity', 'and', 0.85, 1.0)
sort_region (IntermediateBalls, FinalBalls, 'first_point', 'true', 'column')

9） Sort the smallest area

smallest_circle (FinalBalls, Row, Column, Radius)
NumBalls := |Radius|
Diameter := 2*Radius
meanDiameter := sum(Diameter)/NumBalls
mimDiameter := min(Diameter)
dev_display (Bond)

disp_circle (WindowID, Row, Column, Radius)
dev_set_color ('white')
for i := 1 to NumBalls by 1
    if (fmod(i,2)=1)
        disp_message (WindowID, 'D: '+Diameter[i-1], 'image', Row[i-1]-2.7*Radius[i-1], max([Column[i-1]-60,0]), 'white', 'false')
    else
        disp_message (WindowID, 'D: '+Diameter[i-1], 'image', Row[i-1]+1.2*Radius[i-1], max([Column[i-1]-60,0]), 'white', 'false')
    endif
endfor

10） Results output ：