Halcon's region: features of multiple regions (6)

One 、 summary  

         The operator of the region , It's best for beginners to try every one , This is not just a matter of enhancing the image , And understand Halcon The process of ability , And improve the efficiency of engineering practice . This paper introduces several applications in numerical image analysis , More useful domain operators . Please share ！

Two 、 Operators and experiments

2.1  Regional expansion

• expand_region(Regions, ForbiddenArea : RegionExpanded : Iterations, Mode : )

Fill gaps between areas or split overlapping areas .ForbiddenArea Is not involved in the operation （ Retain ） Region .

•   Test code ：

read_image(Image, 'F:/images/maps/manyRegs')
rgb1_to_gray(Image,gray)
get_image_size(Image, Width, Height)
threshold(gray, Region, 0, 215)
connection(Region, ConRegions)
gen_empty_region(EmptyRegion)
expand_region(ConRegions, EmptyRegion, RegionExpanded, 'maximal', 'image')

Results show ：

First, divide the area into multiple groups .

                 

Expand the region of multivariate group , Spread the whole map area . 

 

2.2 Select... From the area tuple index Regions

        select_obj Will have Index Index given （ from 1 Start ） Icon object from icon input object tuple Objects Copy to output object ObjectSelected. No new storage space is allocated for areas and images . contrary , Creates a new object that contains a reference to an existing object . The number of objects in an object tuple can be manipulated count_obj To query .

Such as ： Take out No 12 Individuals of regions , Deposit in  ObjectSelected Tuples

select_obj (RegionExpanded, ObjectSelected,12)

2.3  The nearest neighbors of the region are listed  

      find_neighbors   function ： Search for immediate neighborhood .

• find_neighbors (ObjectSelected, RegionExpanded, 1, RegionIndex1, RegionIndex2)

Output  RegionIndex1, RegionIndex2 Deposit ,ObjectSelected Sequence number pairs of all nearest neighbors of the region .

Reference code ：

read_image(Image, 'F:/images/maps/manyRegs')
rgb1_to_gray(Image,gray)
get_image_size(Image, Width, Height)
threshold(gray, Region, 0, 215)
connection(Region, ConRegions)
gen_empty_region(EmptyRegion)
expand_region(ConRegions, EmptyRegion, RegionExpanded, 'maximal', 'image')

select_obj (RegionExpanded, ObjectSelected,1)
find_neighbors (RegionExpanded, RegionExpanded, 3, RegionIndex1, RegionIndex2)

  3、 ... and 、 An example of an official tutorial

The following is an example from the Official Handbook , The result is to extract the nearest neighbors of all regions .

read_image (Image, 'fabrik')
dev_close_window ()
dev_open_window (0, 0, 512, 512, 'black', WindowID)
dev_set_colored (6)
regiongrowing (Image, Regions, 1, 1, 3, 200)
gen_empty_region (Empty)
expand_region (Regions, Empty, RegionExpanded, 'maximal', 'image')
dev_clear_window ()
select_obj (RegionExpanded, ObjectSelected, 19)
find_neighbors (ObjectSelected, RegionExpanded, 1, RegionIndex1, RegionIndex2)
select_obj (RegionExpanded, ObjectSelected, RegionIndex2)
gen_empty_obj (NoObj)
find_neighbors (NoObj, RegionExpanded, 1, RegionIndex1, RegionIndex2)
shape_trans (RegionExpanded, RegionTrans, 'inner_center')
area_center (RegionTrans, Area, Row, Column)
stop ()
dev_clear_window ()
dev_set_colored (6)
dev_display (RegionExpanded)
dev_set_draw ('margin')
dev_set_color ('black')
dev_set_line_width (1)
dev_display (RegionExpanded)
dev_set_color ('white')
dev_set_line_width (3)
for i := 0 to |RegionIndex1| - 1 by 1
    disp_line (WindowID, Row[RegionIndex1[i] - 1], Column[RegionIndex1[i] - 1], Row[RegionIndex2[i] - 1], Column[RegionIndex2[i] - 1])
endfor

Reference results ：