Scribd
Upload
49 views26 pages

2019-22-Hough Transform

The document discusses the Hough transform, which is a method for detecting shapes within images. It transforms a point in Cartesian space to a curve in parameter space. Accumulator cells in parameter space vote for parameter values corresponding to shapes that pixels may lie on. Peaks in accumulator cells indicate detected shapes. The Hough transform can be used for edge linking and boundary detection in images by finding high concentrations of pixels in accumulator cells.

Uploaded by

Renjeesh bkl
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
49 views26 pages

2019-22-Hough Transform

The document discusses the Hough transform, which is a method for detecting shapes within images. It transforms a point in Cartesian space to a curve in parameter space. Accumulator cells in parameter space vote for parameter values corresponding to shapes that pixels may lie on. Peaks in accumulator cells indicate detected shapes. The Hough transform can be used for edge linking and boundary detection in images by finding high concentrations of pixels in accumulator cells.

Uploaded by

Renjeesh bkl
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 26

References

• Gonzalez and Woods, Digital Image Processing, Pearson Education,3/E, (2008)


Ch-10
• The images in the presentation are taken from text book Gonzales and
Woods and different internet websites

25-Hough Transform
2 7:11 PM
Outcomes

• Edge Linking and Boundary Detection


Global Processing
Hough Transform

25-Hough Transform
3 7:11 PM
Global Processing using Hough Transform

• Previous methods discussed are applicable


to situations where knowledge about pixels belonging to individual
objects is at least partially available

25-Hough Transform
4 7:11 PM
Global Processing using Hough Transform

• An approach based on whether sets of pixels lie on curves of a


specified shape
• Once detected , these curves form the edges or region boundaries
of interest.

25-Hough Transform
5 7:11 PM
Hough Transform

• Given n points in an image .


• We want to find subset of these points that lie on straight lines
• Solution
Find all lines determined by every pair of points
Find all subsets of points that are close to particular lines.
But this is computationally expensive

25-Hough Transform
6 7:11 PM
Hough Transform

• Hough proposed an alternative approach in 1962


• Known as the Hough Transform

25-Hough Transform
7 7:11 PM
Hough Transform
• Consider a point ( ) in the x-y plane

• Slope - intercept form of the equation for str.


line is

25-Hough Transform
8 7:11 PM
Hough Transform

• Infinitely many lines pass through this point

• Satisfying for different values


of a and b

25-Hough Transform
9 7:11 PM
Hough Transform
• Rewriting the equation

as

• Considering the ab plane or


parameter space
yields the equation of a single line for a
fixed pair

25-Hough Transform
10 7:11 PM
Hough Transform
• A second point , also has a line in the
parameter space

• Unless they are parallel, they intersect at ( )


• Where
a’ is the slope
b’-the intercept of the line containing both
( ) and ( ) in the x-y plane

25-Hough Transform
11 7:11 PM
Hough Transform

All the points on this line have lines in


the parameter space that intersect at
(a’,b’)

25-Hough Transform
12 7:11 PM
Hough Transform

Practical difficulty:

• For vertical lines the slope approaches infinity

• Use normal representation of line

25-Hough Transform
13 7:11 PM
Hough Transform
Normal Representation of line

Horz line has θ=0, with ρ : +ve x intercept


Vertical line θ=90, with ρ : +ve y intercept
Or θ=-90, with ρ : -ve y intercept

25-Hough Transform
14 7:11 PM
Hough Transform

• Each sinusoidal curve represents a family of lines


that pass through a particular point ( ) in
the x-y plane

• The intersection point (ρ’, θ’) corresponds to the


line that passes through the points ( ),
( )

25-Hough Transform
15 7:11 PM
Hough Transform
Parameter space is subdivided into accumulator
cells

D is the distance between opposite corners of


the image

25-Hough Transform
16 7:11 PM
Hough Transform

The cell at coordinates with accumulator


value ,
corresponds to the square associated with
parameter space co-ordinates

Initially , these cells are set to zero

25-Hough Transform
17 7:11 PM
Hough Transform
For every non- background point, , in x-y plane,
We let equal each of the allowed subdivision values
on the axis and
Solve for the corresponding , using the equation

The resulting values are rounded off to the nearest


allowed cell value along the axis.

25-Hough Transform
18 7:11 PM
Hough Transform

If

Finally, value P in means that there are


P points that lie on the line

25-Hough Transform
19 7:11 PM
Hough Transform

No of subdivisions in the ρθ plane ,


determines the accuracy of the
collinearity of these points

25-Hough Transform
20 7:11 PM
Hough Transform

Hough Transform is applicable to any function

g(v,c)=0

v-vector coordinates
c-vector coefficients

25-Hough Transform
21 7:11 PM
Hough Transform
Points lying on the circle:

Three parameters

3D parameter space with cube like cells of the form

25-Hough Transform
22 7:11 PM
Hough Transform

Procedure is to choose then find

And update accumulator cells

Complexity depends on the number of co-ordinates and coefficients

25-Hough Transform
23 7:11 PM
Edge Linking using Hough Transform

• Obtain binary edge image using the techniques discussed


• Specify the subdivision in the ρθ plane
• Examine the counts of accumulator cells for high pixel concentration
• Examine the relation ship between pixel in a chosen cell

25-Hough Transform
24 7:11 PM
Conclusions

• Hough Transform

• Edge linking using Hough Transform

25-Hough Transform
25 7:11 PM

You might also like