halcon学习笔记(3)——critical_points例子学习(图像分割Region)
例子在HDevelop自带例子里面，叫critical_points；目的是在所画的闭合区域内的进行检测图像中的鞍点，鞍点具体是什么我这里也不太了解；以后深入再叙述吧。话不多说先上例子：
例子中输入图像和处理后的图像是这样子：
左边是输入图像，白色的圈是鼠标画出的闭合区域，右边是图像分割后图像，并检测鞍点，就是图中黑白交错的圆圈；我使用的图形如下：
实验图片(1)
实验图片(2)
从上图可以大概知道鞍点特异性，当然鞍点不是这次重点，重点是检测给定区域内的项目，忽略区域外的区域；代码具体如下：
dev_update_window ('off')
* ****
* step: acquire image
* ****
read_image (Image, 'C:/Users/shanwenjun/Desktop/22.png')
get_image_size (Image, Width, Height)
dev_close_window ()
dev_open_window (0, 0, Width, Height, 'black', WindowHandle)
dev_set_part (0, 0, Height - 1, Width - 1)
dev_set_line_width (3)
dev_set_color ('blue')
dev_set_draw ('margin')
dev_display (Image)
set_display_font (WindowHandle, 20, 'mono', 'true', 'false')
* ****
* step: draw region
* ****
disp_message (WindowHandle, 'Draw with the mouse the region of interest', 'window', -1, -1, 'blue', 'false')
dev_set_color ('white')
draw_region (Region, WindowHandle)
dev_display (Region)
disp_continue_message (WindowHandle, 'black', 'true')
stop ()
* ****
* step: create ROI
* ****
reduce_domain (Image, Region, ImageReduced)
disp_continue_message (WindowHandle, 'black', 'true')
stop ()
* ****
* step: process image within ROI
* ****
critical_points_sub_pix (ImageReduced, 'facet', 1.5, 8, RowMin, ColumnMin, RowMax, ColumnMax, RowSaddle, ColSaddle)
dev_clear_window ()
dev_display (ImageReduced)
dev_set_color ('yellow')
for i := 0 to |RowSaddle| - 1 by 1
gen_cross_contour_xld (Cross, RowSaddle[i], ColSaddle[i], 25, 0.785398)
dev_display (Cross)
endfor
stop ()
dev_update_window ('on')这里还是一句一句来：
dev_update_window ('off')
* ****
* step: acquire image
* ****
read_image (Image, 'C:/Users/shanwenjun/Desktop/22.png')
get_image_size (Image, Width, Height)
dev_close_window ()
dev_open_window (0, 0, Width, Height, 'black', WindowHandle)
dev_set_part (0, 0, Height - 1, Width - 1)
dev_set_line_width (3)
dev_set_color ('blue')
dev_set_draw ('margin')
dev_display (Image)
set_display_font (WindowHandle, 20, 'mono', 'true', 'false')
* ****
* step: draw region
* ****
disp_message (WindowHandle, 'Draw with the mouse the region of interest', 'window', -1, -1, 'blue', 'false')dev_set_color ('white')*设置颜色； draw_region (Region, WindowHandle)*设置Region画笔dev_display (Region)*显示所画的区域 disp_continue_message (WindowHandle, 'black', 'true')*显示一句话：按F5继续stop ()*暂停等待按F5reduce_domain (Image, Region, ImageReduced)*剪切只显示Region区域的图像disp_continue_message (WindowHandle, 'black', 'true')*显示一句话：按F5继续stop ()*暂停等待按F5critical_points_sub_pix (ImageReduced, 'facet', 1.5, 8, RowMin, ColumnMin, RowMax, ColumnMax, RowSaddle, ColSaddle)*检测鞍点的函数，最后两个参数RowSaddle和ColSaddle是鞍点的xy坐标数组，前面的查看帮助文档吧，这个函数后面具体应该还会用到，原理现*在不太清楚dev_clear_window ()*清空窗口dev_display (ImageReduced)*显示剪切后图像dev_set_color ('yellow')*设置颜色for i := 0 to |RowSaddle| - 1 by 1*循环读取鞍点
gen_cross_contour_xld (Cross, RowSaddle[i], ColSaddle[i], 25, 0.785398)*叉子标记，用鞍点的坐标dev_display (Cross)*显示叉子endfor*结束循环stop ()*停止dev_update_window ('on')*停止更新窗口这节重点要学习是上面的Region语句，图像分割，在图像检测中非常重要，因为可以缩小检测范围，提高检测速度和准确率，此次的Region只学习例子程序，Region还要在深入，比如画后的如何保存，每次都可在这个范围内检测等等。