博主寫作不容易，孩子需要您鼓勵 萬水千山總是情 , 先點個贊行不行

處理要求

橢圓/圓環（產品易變形，為橢圓）內外圓毛刺（凸起）缺口（凹陷）檢測。

處理方法1

方法一思路

1、這是一個圓環產品檢測，我們可以通過產品區域與標準圓環進行比較得出不良區域。
2、為了避免誤檢、誤判，我們可以通過區域篩選閾值偏移的方法濾除干擾區域，可以將標準圓環放大消除一些圓度導致干擾。
3、根據不同用戶的精度要求，可以通過調節缺陷面積進行篩選。
4、方法1的代碼量有點多，但是更貼近工業現場使用。

方法一halcon源碼

dev_close_window ()
read_image (Image, 'C:/Users/22967/Desktop/圓環缺陷檢測/處理1.jpg')
dev_open_window_fit_image (Image, 0, 0, -1, -1, WindowHandle)
dev_display (Image)*********************方法一**************************
*****變量定義
* 卡尺測量參數
CenterRow:=0
CenterColumn:=0
CenterRadius:=0
* 灰度分割閾值偏移
ThresholdOffest:=80
* 缺陷區域面積閾值
NGArea:=50
*圓環內外偏移閾值
RadiusOffest:=5
* Image Acquisition 01: Code generated by Image Acquisition 01
list_files ('C:/Users/22967/Desktop/圓環缺陷檢測', ['files','follow_links'], ImageFiles)
tuple_regexp_select (ImageFiles, ['\\.(tif|tiff|gif|bmp|jpg|jpeg|jp2|png|pcx|pgm|ppm|pbm|xwd|ima|hobj)$','ignore_case'], ImageFiles)
for Index := 0 to |ImageFiles| - 1 by 1read_image (Image, ImageFiles[Index])rgb1_to_gray (Image, GrayImage)*****圓環灰度篩選binary_threshold (GrayImage, Region, 'max_separability', 'dark', UsedThreshold)threshold (GrayImage, Region1, 0, UsedThreshold+ThresholdOffest)*分割連通域connection (Region1, ConnectedRegions)*選取圓環區域select_shape_std (ConnectedRegions, SelectedRegions, 'max_area', 70)*濾除圓環邊緣毛刺opening_circle (SelectedRegions, RegionOpening, 1.5)*****求圓環內外圓*求圓環外圓smallest_circle (RegionOpening, Row, Column, Radius)CenterRow[0]:=RowCenterColumn[0]:=ColumnCenterRadius[0]:=Radius*求圓環內圓fill_up (RegionOpening, RegionFillUp)difference (RegionFillUp, RegionOpening, RegionDifference)connection (RegionDifference, ConnectedRegions2)select_shape_std (ConnectedRegions2, SelectedRegions1, 'max_area', 70)smallest_circle (SelectedRegions1, Row1, Column1, Radius1)CenterRow[1]:=Row1CenterColumn[1]:=Column1CenterRadius[1]:=Radius1*****對內外圓進行卡尺測量*創建測量句柄create_metrology_model (MetrologyHandle)*設置卡尺測量參數add_metrology_object_circle_measure (MetrologyHandle, CenterRow, CenterColumn, CenterRadius, CenterRadius[0]/10, CenterRadius[0]/60, 1, 4, ['measure_distance','min_score'], [CenterRadius[0]/30,0.2], Indexnumb)*進行測量apply_metrology_model (Image, MetrologyHandle)*得到測量結果get_metrology_object_result (MetrologyHandle, 'all', 'all', 'result_type', 'all_param', Parameter)get_metrology_object_result_contour (Contour, MetrologyHandle, 'all', 'all', 1.5)get_metrology_object_measures (Contours, MetrologyHandle, 'all', 'all', Row1, Column1)*****求出標準圓環,進行缺陷檢測*突出部分gen_circle (Circle, Parameter[0], Parameter[1], Parameter[2]+RadiusOffest)gen_circle (Circle1, Parameter[3], Parameter[4], Parameter[5]-RadiusOffest)difference (Circle, Circle1, RegionDifference1)difference (SelectedRegions, RegionDifference1, RegionDifference2)*內凹部分gen_circle (Circle2, Parameter[0], Parameter[1], Parameter[2]-RadiusOffest)gen_circle (Circle3,Parameter[3], Parameter[4], Parameter[5]+RadiusOffest)difference (Circle2, Circle3, RegionDifference4)difference (RegionDifference4, SelectedRegions, RegionDifference3)*濾除噪點opening_circle (RegionDifference2, RegionOpening1, 1.5)opening_circle (RegionDifference3, RegionOpening2, 1.5)*合并缺陷區域union2 (RegionOpening1, RegionOpening2, RegionUnion)closing_circle (RegionUnion, RegionClosing, 3.5)connection (RegionClosing, ConnectedRegions1)*結果判斷area_center (ConnectedRegions1, Area, Row2, Column2)count_obj (ConnectedRegions1, Number)gen_empty_obj (EmptyObject)for Index1 := 1 to Number by 1if  (Area[Index1-1] > NGArea)select_obj (ConnectedRegions1, ObjectSelected, Index1)smallest_circle (ObjectSelected, Row3, Column3, Radius2)gen_circle (Circle4, Row3, Column3, Radius2)concat_obj (EmptyObject, Circle4, EmptyObject)endifendfordev_set_draw ('margin')dev_set_line_width (3)dev_display (Image)dev_display (EmptyObject)
*     stop()
endfor
clear_metrology_model (MetrologyHandle)

處理效果

處理方法2

方法二思路

1、利用形態學方法進行缺陷檢測。
2、缺點就是對圓度不敏感。

方法二halcon源碼

dev_close_window ()
read_image (Image, 'C:/Users/22967/Desktop/圓環缺陷檢測/處理1.jpg')
dev_open_window_fit_image (Image, 0, 0, -1, -1, WindowHandle)
dev_display (Image)
*********************方法二**************************
* 灰度分割閾值偏移
ThresholdOffest:=80
*外圓缺陷查找閾值
OutCircleTh:=200.5
*內圓缺陷查找閾值
InCircleTh:=100.5
*缺陷區域面積閾值
NGArea:=50
*噪點過濾閾值
DelNoise:=1.5* Image Acquisition 01: Code generated by Image Acquisition 01
list_files ('C:/Users/22967/Desktop/圓環缺陷檢測', ['files','follow_links'], ImageFiles)
tuple_regexp_select (ImageFiles, ['\\.(tif|tiff|gif|bmp|jpg|jpeg|jp2|png|pcx|pgm|ppm|pbm|xwd|ima|hobj)$','ignore_case'], ImageFiles)
for Index := 0 to |ImageFiles| - 1 by 1*讀入圖片read_image (Image, ImageFiles[Index])rgb1_to_gray (Image, GrayImage)*二值化選取墊片區域binary_threshold (GrayImage, Region, 'max_separability', 'dark', UsedThreshold)threshold (GrayImage, Region1, 0, UsedThreshold+ThresholdOffest)connection (Region1, ConnectedRegions)select_shape_std (ConnectedRegions, SelectedRegions, 'max_area', 70)*外圓缺陷查找fill_up (SelectedRegions, RegionFillUp1)opening_circle (RegionFillUp1, RegionOpening, OutCircleTh)difference (RegionFillUp1, RegionOpening, RegionDifference5)*內圓缺陷查找difference (RegionFillUp1, SelectedRegions, RegionDifference6)connection (RegionDifference6, ConnectedRegions3)select_shape_std (ConnectedRegions3, SelectedRegions2, 'max_area', 70)opening_circle (SelectedRegions2, RegionOpening3, InCircleTh)difference (SelectedRegions2, RegionOpening3, RegionDifference7)*合并缺陷區域union2 (RegionDifference5, RegionDifference7, RegionUnion1)opening_circle (RegionUnion1, RegionOpening4, DelNoise)connection (RegionOpening4, ConnectedRegions4)*結果判斷area_center (ConnectedRegions4, Area1, Row4, Column4)gen_empty_obj (EmptyObject1)for Index1 := 1 to |Area1| by 1if  (Area1[Index1-1] > NGArea)select_obj (ConnectedRegions4, ObjectSelected, Index1)smallest_circle (ObjectSelected, Row3, Column3, Radius2)gen_circle (Circle4, Row3, Column3, Radius2)concat_obj (EmptyObject1, Circle4, EmptyObject1)endifendfor  *顯示結果dev_set_draw ('margin')dev_set_line_width (3)dev_display (Image)dev_display (EmptyObject1)stop()
endfor

處理效果

------------------------------------------------------------------------------------------分割線
------------------------------------------------------------------------------------------

博主簡介： 工業自動化上位機軟件工程師、機器視覺算法工程師、運動控制算法工程師。目前從業于智能制造自動化行業。 博主郵箱：2296776525@qq.com

幫忙點個贊吧。哈哈。