毛刺检测

王俊杰

边缘毛刺检测，求教各位大佬，有没有好的解决方案

z7172930

dev_update_off()
dev_set_colored(12)
dev_set_draw('fill')
read_image (Image, 'C:/Users/Bells/Desktop/154921zud2edj04v1zkazj.bmp')
threshold(Image, Region, 150, 255)
opening_circle(Region, RegionOpening, 2.5)
closing_circle(RegionOpening, RegionClosing, 3.5)
connection(RegionClosing, ConnectedRegions)
select_shape(ConnectedRegions, SelectedRegions, ['width','area'], 'and', [20,1000], [999,99999])
dev_display(Image)
dev_display(SelectedRegions)

王俊杰

z7172930 发表于 2022-5-11 17:27
dev_update_off()
dev_set_colored(12)
dev_set_draw('fill')

那个不属于毛刺，绿色区域旁边的那个才是毛刺，绿色区域的是胶体

z7172930

那就建议重新打光

王俊杰

z7172930 发表于 2022-5-12 13:38
那就建议重新打光

产品材质的问题，这是目前最好的效果了

z7172930

方法1 AOI光源,彩色相机
方法2 光立度打光

qq27472007

毛刺和胶体成像效果无法区分, 那可就很难了

glorialantian

需要把毛刺和胶体通过打光区分开，然后再图像处理，你现在用的黑白相机？

余跃

read_image (Image, 'C:/Users/47863/Desktop/图片/9.bmp')
dev_close_window ()
get_image_size (Image, Width, Height)
dev_open_window (0, 0, Width/2, Height/2, 'black', WindowHandle)
dev_display (Image)
threshold (Image, Regions, 72, 255)
connection (Regions, ConnectedRegions)
select_shape (ConnectedRegions, SelectedRegions, 'area', 'and', 9000, 11000)
fill_up (SelectedRegions, RegionFillUp)
opening_circle (RegionFillUp, RegionOpening, 3.5)
connection (RegionOpening, ConnectedRegions1)
select_shape (ConnectedRegions1, SelectedRegions1, 'area', 'and', 2500, 3500)
shape_trans (SelectedRegions1, RegionTrans, 'rectangle1')
shape_trans (SelectedRegions1, RegionTrans1, 'inner_rectangle1')
difference (RegionTrans, RegionTrans1, RegionDifference)
intersection (SelectedRegions1, RegionDifference, RegionIntersection)
opening_circle (RegionIntersection, RegionOpening1, 2)
connection (RegionOpening1, ConnectedRegions2)
select_shape (ConnectedRegions2, SelectedRegions2, 'area', 'and', 500, 1000)
reduce_domain (Image, SelectedRegions2, ImageReduced)
threshold (ImageReduced, Regions1, 0, 255)
dev_display (Image)
dev_display (ImageReduced)
dev_display (Regions1)

glorialantian

余跃 发表于 2022-5-15 23:34
read_image (Image, 'C:/Users/47863/Desktop/图片/9.bmp')
dev_close_window ()
get_image_size (Image, W ...

按照你的方法，做的结果，非常赞