Halcon每日一练

## 灰度较低的背景下提取轮廓线

要求：

低灰度下边缘提取

图片：

*Author：ygw
*Date:2021/06/29
***********************功能：低灰度下边缘提取*************************
*思路：
*1.因灰度较低无法正常用单阈值进行分割，采取锐化图像获取较明显边缘
*2.提取黑白部分并进行预处理
*3.形状转换获得较好的区域形状输出

***********************************程序主体*************************************
dev_update_off ()
dev_close_window ()
read_image (Image, 'F:/halcon学习/学习-halcon/1-算法入门作业培训阶段1/2/原图.jpg')
get_image_size (Image, Width, Height)
rgb1_to_gray (Image, GrayImage)
dev_open_window_fit_image (GrayImage, 0, 0, -1, -1, WindowHandle)
dev_display (GrayImage)
mean_image (GrayImage, ImageMean, Width/10, Height/10)
*sub_image (Image, ImageMean, ImageSub, 50, 80)对两幅图像做减法
 *ImageSub=(Image-ImageMean)*Mult+Add
 *这里为了得到锐化图像凸显较为明显的边缘
 *算子中Image为原图，ImageMean为滤波后的图像，乘以Mult是要拉大两幅图像相减后的对比度，加上一个Add是为了增加相减后的图像的整体亮度
sub_image (GrayImage, ImageMean, ImageSub, 50, 80)
dev_display (ImageSub)
*对得到的锐化图像进行进一步的滤波去噪点
mean_image (ImageSub, ImageMean1, 10, 10)
*获取黑色、白色部分
threshold (ImageMean1, Region1, 230, 255)
threshold (ImageMean1, Region2, 0, 10)
*将region2的区域合并到region1中
union2 (Region1, Region2, RegionUnion)
connection (RegionUnion, ConnectedRegions)
select_shape (ConnectedRegions, SelectedRegions, 'area', 'and', 100, 99999)
*将所有区域合并为一个区域
union1 (SelectedRegions, RegionUnion1)
closing_circle (RegionUnion1, RegionClosing, 50)
connection (RegionClosing, ConnectedRegions1)
select_shape_std (ConnectedRegions1, SelectedRegions1, 'max_area', 70)
fill_up (SelectedRegions1, RegionFillUp)
difference (RegionFillUp, SelectedRegions1, RegionDifference)
shape_trans (RegionDifference, RegionTrans1, 'convex')
shape_trans (SelectedRegions1, RegionTrans2, 'convex')
difference (RegionTrans2, RegionTrans1, RegionRing) 
dev_set_draw ('margin')
dev_display (Image)
dev_set_color ('blue')
dev_set_line_width (2)
dev_display (RegionRing)

效果