Defect detection, introduction to Halcon case.

defect detection ,halcon Introduction to cases .
Common defects ：
1. Convex concave structure （ Contains small burrs ）.
2. Internal stain , Uneven surface , defects , Holes , be damaged , burns , Youze .
3. Scratches .
processing method ：

1. Polishing treatment
   Scratches . Low angle ring light , Turn the background white , The scratches darken .
   

Bump , Use a low angle ring light , Darken the background , The convex point is whitened .
Pit , Use vertical light （ Can be annular ）, Light up the background , The background is darkened .
【 You can use a light bar to fight 】
Surface Fonts （ Bulge ）, Use tile mounted curved light plate , Drive into the side at an angle , The font can be highlighted .
2. Algorithm to deal with
Blob analysis + feature detection
Blob analysis + features + Difference
frequency + Space
Photometric stereometry
Feature training （ classifier , Deep learning ）
measurement + fitting
3. halcon Defect cases .
Case a ： routine ：blob analysis ——check_hazelnut_wafers.hdev（ Check the cookies ）
sketch ：inspect quality of hazelnut wafers( Check the quality of biscuits )
// The surface is damaged in a large area
Blob+ feature detection

1. Take images
2. Binary segmentation binary_threshold
   binary_threshold (Image, Foreground, ‘smooth_histo’, ‘light’, UsedThreshold)
   // Image Read in the picture ,
   // Foreground Output area
   //'smooth_histo’ Segmentation algorithm
   // 'light’ Extract the bright part
   // UsedThreshold Output parameters
3. Morphological open operation opening_circle（ Remove a small number of pixels ）
4. feature extraction area_holes ( Find the hole )
5. Directly calculate the rectangularity rectangularity
6. Use interpretation if. Judge the value of holes or rectangularity
   Case 2 ： routine ：blob analysis ——fin.hdev（ Finned object ）
   sketch ：inspect a contour for fins.（ Check the outline of the heat sink ）
   //
   Blob analysis + Difference + feature detection
7. Take images
8. Binary segmentation binary_threshold
   binary_threshold (Image, Background, ‘max_separability’, ‘light’, UsedThreshold)
   // 'max_separability’ Maximum separation method
9. Closed operation closing_circle( Fill area )
10. Difference operation difference
11. Morphological open operation
12. Find the area
13. Judge whether it is a defect

Case three ： routine ：blob analysis ——inspect_bottle_mouth.hdev（ Check the bottle mouth ）
sketch ：check bottle mouths for defects（ Check whether the bottle mouth is defective ）
Blob+ feature detection
Case four ： routine ：blob analysis ——novelty_detection_dyn_threshold.hdev
sketch ：inspect a web using dyn_threshold（ Use dyn_ Threshold check network ）
// Used to stabilize the light , The environment is simple
Code ：
dev_update_window (‘off’)
read_image (Image, ‘plastic_mesh/plastic_mesh_01’)
dev_close_window ()
get_image_size (Image, Width, Height)
dev_open_window_fit_image (Image, 0, 0, Width, Height, WindowHandle)
set_display_font (WindowHandle, 18, ‘mono’, ‘true’, ‘false’)
dev_set_draw (‘margin’)
dev_set_line_width (3)
for J := 1 to 14 by 1
read_image (Image, ‘plastic_mesh/plastic_mesh_’ + J$‘02’)
mean_image (Image, ImageMean, 49, 49)
// Mean filtering , Pull the grayscale value to smooth
dyn_threshold (Image, ImageMean, RegionDynThresh, 5, ‘dark’)
// Use local threshold to segment the image ,
// Image： The input image .ImageMean： Enter the comparison image .RegionDynThresh： Output image .5,： threshold .‘dark’： Select the dark part
// Establish a detection system for image gray value .5： Is the threshold . Calculate the input image and the comparison image , You can select the bright and dark parts of the input image .‘dark’： To get the dark part .‘light’： Select the bright part for .Lightdark： Select the original part for
// It is often used after mean filtering
connection (RegionDynThresh, ConnectedRegions)
// feature extraction
select_shape (ConnectedRegions, ErrorRegions, ‘area’, ‘and’, 500, 99999)
count_obj (ErrorRegions, NumErrors)
dev_display (Image)
dev_set_color (‘red’)
dev_display (ErrorRegions)
if (NumErrors > 0)
disp_message (WindowHandle, ‘Mesh not OK’, ‘window’, 24, 12, ‘black’, ‘true’)
else
disp_message (WindowHandle, ‘Mesh OK’, ‘window’, 24, 12, ‘black’, ‘true’)
endif
if (J < 14)
disp_continue_message (WindowHandle, ‘black’, ‘true’)
stop ()
endif
endfor

Case 5 ： routine ：blob analysis ——novelty_detection_dyn_threshold.hdev
sketch ：inspect a web using dyn_threshold（ Use dyn_ Threshold check network ）
// be used for pcb Detection of small line breaks
Code

*blob+ Difference + features
read_image (Image, ‘pcb’)
dev_close_window ()
get_image_size (Image, Width, Height)
dev_open_window (0, 0, Width, Height, ‘black’, WindowHandle)
dev_display (Image)
// Grayscale morphology
// Grayscale open operation , corrosion , Dark pixels increase
gray_opening_shape (Image, ImageOpening, 7, 7, ‘octagon’)
// Gray closed operation , inflation , Bright pixels increase
gray_closing_shape (Image, ImageClosing, 7, 7, ‘octagon’)
// Difference
dyn_threshold (ImageOpening, ImageClosing, RegionDynThresh, 75, ‘not_equal’)
dev_display (Image)
dev_set_color (‘red’)
dev_set_draw (‘margin’)
dev_display (RegionDynThresh)

Case 6 ： routine ：blob analysis ——check_blister.hdev（ Test a tablet ）
// Test a tablet for defects .

Code

• location ,blob analysis + Feature recognition

dev_close_window ()
dev_update_off ()
read_image (ImageOrig, ‘blister/blister_reference’)
dev_open_window_fit_image (ImageOrig, 0, 0, -1, -1, WindowHandle)
set_display_font (WindowHandle, 14, ‘mono’, ‘true’, ‘false’)
dev_set_draw (‘margin’)
dev_set_line_width (3)
access_channel (ImageOrig, Image1, 1)
// Equivalent to the gray conversion effect
// A channel that accesses a multi-channel image . That's right RGB3 One channel of the channel is extracted , Equivalent to converting to grayscale image
threshold (Image1, Region, 90, 255)
// Two valued
shape_trans (Region, Blister, ‘convex’)
orientation_region (Blister, Phi)
// Detection angle
area_center (Blister, Area1, Row, Column)
vector_angle_to_rigid (Row, Column, Phi, Row, Column, 0, HomMat2D)
affine_trans_image (ImageOrig, Image2, HomMat2D, ‘constant’, ‘false’)
// Affine variation
gen_empty_obj (Chambers)
// Count
for I := 0 to 4 by 1
Row := 88 + I * 70
for J := 0 to 2 by 1
Column := 163 + J * 150
gen_rectangle2 (Rectangle, Row, Column, 0, 64, 30)
concat_obj (Chambers, Rectangle, Chambers)
// Display, draw image and count the single area of the tablet
// Put the extracted image of a single tablet on Chambers Array , Facilitate subsequent de intersection
endfor
endfor
affine_trans_region (Blister, Blister, HomMat2D, ‘nearest_neighbor’)
// Affine change the region
difference (Blister, Chambers, Pattern)
// Find the difference . Calculate the difference between the affine transformation region and the tablet region , Used to extract tablet area
union1 (Chambers, ChambersUnion)
// Form a connected domain
orientation_region (Blister, PhiRef)
// Calculate the angle of the positive area
PhiRef := rad(180) + PhiRef
area_center (Blister, Area2, RowRef, ColumnRef)
// Find the coordinates
// End of positioning
// defect detection （ Find the intersection between the tablet and the original area at the standard position ）
Count := 6
for Index := 1 to Count by 1
read_image (Image, ‘blister/blister_’ + Index$‘02’)
threshold (Image, Region, 90, 255)
connection (Region, ConnectedRegions)
select_shape (ConnectedRegions, SelectedRegions, ‘area’, ‘and’, 5000, 9999999)
shape_trans (SelectedRegions, RegionTrans, ‘convex’)
orientation_region (RegionTrans, Phi)
area_center (RegionTrans, Area3, Row, Column)
vector_angle_to_rigid (Row, Column, Phi, RowRef, ColumnRef, PhiRef, HomMat2D)
affine_trans_image (Image, ImageAffineTrans, HomMat2D, ‘constant’, ‘false’)
// Find the area where the image is located , It is convenient to crop the image

// Tablet segmentation
reduce_domain (ImageAffineTrans, ChambersUnion, ImageReduced)
// Crop the tablet area of the image
decompose3 (ImageReduced, ImageR, ImageG, ImageB)
//hsv conversion
var_threshold (ImageB, Region, 7, 7, 0.2, 2, ‘dark’)
// Yes v To binarize
connection (Region, ConnectedRegions0)
// To break off
closing_rectangle1 (ConnectedRegions0, ConnectedRegions, 3, 3)
// Closed operation
fill_up (ConnectedRegions, RegionFillUp)
// fill
select_shape (RegionFillUp, SelectedRegions, ‘area’, ‘and’, 1000, 99999)
opening_circle (SelectedRegions, RegionOpening, 4.5)
connection (RegionOpening, ConnectedRegions)
select_shape (ConnectedRegions, SelectedRegions, ‘area’, ‘and’, 1000, 99999)
shape_trans (SelectedRegions, Pills, ‘convex’)
*
* Classify segmentation results and display statistics
count_obj (Chambers, Number)
gen_empty_obj (WrongPill)
gen_empty_obj (MissingPill)
// establish 2 An empty count array
for I := 1 to Number by 1
select_obj (Chambers, Chamber, I)
intersection (Chamber, Pills, Pill)
// Take the intersection of the tablet and the area where the tablet is located
area_center (Pill, Area, Row1, Column1)
if (Area > 0)
min_max_gray (Pill, ImageB, 0, Min, Max, Range)
// Find the minimum and maximum gray value
if (Area < 3800 or Min < 60)
concat_obj (WrongPill, Pill, WrongPill)
endif
else
concat_obj (MissingPill, Chamber, MissingPill)
endif
endfor
*
dev_clear_window ()
dev_display (ImageAffineTrans)
dev_set_color (‘forest green’)
count_obj (Pills, NumberP)
count_obj (WrongPill, NumberWP)
count_obj (MissingPill, NumberMP)
dev_display (Pills)
if (NumberMP > 0 or NumberWP > 0)
disp_message (WindowHandle, ‘Not OK’, ‘window’, 12, 12 + 600, ‘red’, ‘true’)
else
disp_message (WindowHandle, ‘OK’, ‘window’, 12, 12 + 600, ‘forest green’, ‘true’)
endif
*
Message := '# Correct pills: ’ + (NumberP - NumberWP)
Message[1] := '# Wrong pills : ’ + NumberWP
Message[2] := '# Missing pills: ’ + NumberMP
*
Colors := gen_tuple_const(3,‘black’)
if (NumberWP > 0)
Colors[1] := ‘red’
endif
if (NumberMP > 0)
Colors[2] := ‘red’
endif
disp_message (WindowHandle, Message, ‘window’, 12, 12, Colors, ‘true’)
dev_set_color (‘red’)
dev_display (WrongPill)
dev_display (MissingPill)
if (Index < Count)
disp_continue_message (WindowHandle, ‘black’, ‘true’)
endif
stop ()
endfor

Case seven ： routine ： Template matching （ Based on relevance ）——measure_fill_level.hdev（ Check the liquid level ）
// Check whether the bottled liquid is too much or too little .
Code ：
// Positioning first , Then set a fixed height , Continue to measure the height
// Take images
dev_close_window ()
dev_update_off ()
read_image (Image, ‘ampoules/ampoules_01’)
get_image_size (Image, Width, Height)
dev_open_window_fit_image (Image, 0, 0, -1, -1, WindowHandle)
dev_set_line_width (2)
dev_set_draw (‘margin’)
set_display_font (WindowHandle, 16, ‘mono’, ‘true’, ‘false’)
// Template matching
// Cutout
gen_rectangle1 (Rectangle, 230, 280, 317, 330)
reduce_domain (Image, Rectangle, ImageModel)
create_shape_model (ImageModel, ‘auto’, 0, 0, ‘auto’, ‘auto’, ‘use_polarity’, ‘auto’, ‘auto’, ModelID)
// Create a template
gen_measure_rectangle2 (0, 0, rad(90), 75, 20, Width, Height, ‘bilinear’, MeasureHandle)
// Form a measurement rectangle
Tolerance := 15
NumImages := 8
for Index := 1 to NumImages by 1
read_image (Image, ‘ampoules/ampoules_’ + Index$’.2d’)
ColumnEdges := []
FillLevelHeight := []
// Define an array
find_shape_model (Image, ModelID, 0, 0, 0.7, 0, 0.1, ‘least_squares’, 0, 0.9, Row, Column, Angle, Score)
// Find the template coordinates
MeanRow := mean(Row)
// Take the mean value of the row coordinates
RefLevel := MeanRow - 160
* Display tolerance area
dev_display (Image)
dev_set_line_width (1)
dev_set_color (‘white’)
gen_rectangle2 (AcceptLevel, RefLevel, mean(Column), 0, 30 + (max(Column) - min(Column)) / 2, Tolerance)
// Form reference rectangle
dev_display (AcceptLevel)
dev_set_line_width (2)
Errors := 0
for Idx := 0 to |Score| - 1 by 1
translate_measure (MeasureHandle, MeanRow - 135, Column[Idx])
// Measure rectangle parallel
measure_pos (Image, MeasureHandle, 2, 7, ‘all’, ‘first’, RowEdge, ColumnEdge, Amplitude, Distance)
// Find the measuring edge
// Continue to subtract the edge of the liquid level from the reference line, which is greater than or less than a certain value to detect whether there is an error .
FillLevelHeight := [FillLevelHeight,RowEdge]
ColumnEdges := [ColumnEdges,ColumnEdge]
gen_cross_contour_xld (Cross, RowEdge, ColumnEdge, 15, 0)
// formation x Mark
gen_rectangle2 (FillLevel, RowEdge, ColumnEdge, 0, 28, 20)
// Form a rectangle
if (abs(FillLevelHeight[Idx] - RefLevel) >= Tolerance)
gen_rectangle2 (ChamberSingle, MeanRow - 133, Column[Idx], 0, 35, 90)
gen_cross_contour_xld (Cross, FillLevelHeight[Idx], ColumnEdges[Idx], 15, 0)
gen_rectangle2 (FillLevel, FillLevelHeight[Idx], ColumnEdges[Idx], 0, 28, 20)
// Draw the wrong area
Errors := Errors + 1
dev_set_color (‘red’)
dev_display (ChamberSingle)
disp_message (WindowHandle, ‘NG’, ‘image’, FillLevelHeight[Idx] - 50, ColumnEdges[Idx] - 10, ‘red’, ‘false’)
else
disp_message (WindowHandle, ‘OK’, ‘image’, FillLevelHeight[Idx] - 50, ColumnEdges[Idx] - 10, ‘green’, ‘false’)
dev_set_color (‘green’)
endif
dev_display (FillLevel)
dev_display (Cross)
endfor
if (Errors > 0)
disp_message (WindowHandle, Errors + ’ BAD’, ‘window’, 10, 12, ‘red’, ‘true’)
else
disp_message (WindowHandle, ‘All OK’, ‘window’, 10, 12, ‘forest green’, ‘true’)
endif
if (Index < NumImages)
disp_continue_message (WindowHandle, ‘black’, ‘true’)
stop ()
endif
endfor