Digital Image Processing, 2nd ed.

27 downloads 2450 Views 2MB Size Report
2002 R. C. Gonzalez & R. E. Woods. Chapter 3: Spatial Processing: NEIGHBORHOOD OF (x,y). Masks or filters operate on neighborhoods of input image ...
Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Chapter 3: Image Enhancement: INTRODUCTION I- GOALS OF IMAGE ENHANCEMENT • Better visual quality for human analysis • Facilitate further image processing

II- SPATIAL DOMAIN PROCESSING •Processing directly the pixels of the image •Single pixel or neighborhood of pixels is processed •Statistical model of gray levels used •Linear or Nonlinear types of processing •Corresponding frequency domain interpretation is possible III- FOURIER/FREQUENCY DOMAIN PROCESSING •Explicitly process the image in the frequency domain (DFT) © 2002 R. C. Gonzalez & R. E. Woods

Digital Image Processing, 2nd ed.

Chapter 3: Spatial Processing: NEIGHBORHOOD OF (x,y)

Masks or filters operate on neighborhoods of input image

© 2002 R. C. Gonzalez & R. E. Woods

www.imageprocessingbook.com

Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Chapter 3: IMAGE AVERAGING (set-up) •Problem: Image has additive noise •Advantage: Multiple images are available (assumed to be aligned) •Noise mode: zero-mean, uncorrelated •Recall: Variance is reduced

g i ( x, y )  f ( x, y )   i ( x, y ) 1 g ( x, y )  K

K

 g ( x, y ) i 1

i

E{g ( x, y )}  f ( x, y ) ;  © 2002 R. C. Gonzalez & R. E. Woods

2 g ( x, y )

1 2    ( x, y ) K

Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Chapter 3: IMAGE AVERAGING (Example)

Simulated Example: Add Gaussian Noise (Std=60) Orig., Noisy Image Average of 8, 16 images Average of 64, 128 images

© 2002 R. C. Gonzalez & R. E. Woods

Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Chapter 3:IMAGE AVERAGING (differences)

Diff. = Original - Averaged (8, 16, 64, 128 images) NOTE: non-zero mean due to negative noise values and scaling for display (?)

© 2002 R. C. Gonzalez & R. E. Woods

Digital Image Processing, 2nd ed. Chapter 3: SPATIAL DOMAIN FILTERING

www.imageprocessingbook.com

Mask Image neighborhood “under” the mask

© 2002 R. C. Gonzalez & R. E. Woods

Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Chapter 3:MASK COEFFICIENTS. and CONVOLUTION EQUATION

t a

g ( x, y ) 

b

  w(s, t ) f ( x  s, y  t )

s  a t b



a

b

  w(s,t ) f ( x  s, y  t )

s  a t b

x  0,1,..., M  1 ; y  0.1,2,...N  1 © 2002 R. C. Gonzalez & R. E. Woods

s

Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Chapter 3: Example Filter Masks, Border Effects

a

Note :

b

  w(s, t ) 1

s  a t b

Implementation Issue: (x,y) near borders, alternatives: Avoid non-overlap, zero-extension, constant extension, replicate rows & columns, symmetric extension, etc. © 2002 R. C. Gonzalez & R. E. Woods

Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Chapter 3: Smoothing Filter Masks for Weighted Averaging

g ( x, y) 

a

b

  w(s, t ) f ( x  s, y  t )

s  a t b

x  0,1,..., M  1 ; y  0.1,2,...N  1 © 2002 R. C. Gonzalez & R. E. Woods

Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Chapter 3: EXAMPLE OF SMOOTHING/AVERAGING

Filter Mask Sizes Used: 3x3, 5x5, 9x9, 15x15, 35x35

© 2002 R. C. Gonzalez & R. E. Woods

Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Chapter 3: Use of Spatial Smoothing before thresholding

•Gross detail is preserved •Apply threshold of 25% of maximum gray level •Small objects do not survive the processing © 2002 R. C. Gonzalez & R. E. Woods

Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Implementing neigborghood operations – MATLAB example % assume a 3x3 mask and anMxM image % zero pad extension of image Xe = X; Xe = [ zeros(1,M); X; zeros(1,M)]; Xe = [zeros(N+2,1) Xe zeros(N+2,1)]; %output image Ye = zeros(size(Xe); %for loops to move over the image for n = 2:N+1 for m = 2:M+1 %for loops to implement 2-D mask new_value = 0; for i = -1:1 for j=-1:1 new_value = new_value + mask(i+2,j+2)*Xe( n+i, m+j) ; end end Ye(n,m) = new_value; end end %remove padding Y = Ye(2:N+1,2:M+1);

© 2002 R. C. Gonzalez & R. E. Woods

Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Chapter 3: Order-Statistics, Median Filters

g ( x, y )  median { f ( x  s, y  t ) : s  [a, a], t  [b, b]} x  0,1,..., M  1 ; y  0,1,2,...N  1

© 2002 R. C. Gonzalez & R. E. Woods

Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Chapter 3: Order-Statistics, Median Filters - II

g ( x, y )  median { f ( x  s, y  t ) : s  [a, a], t  [b, b]} x  0,1,..., M  1 ; y  0,1,2,...N  1 Properties of Median Filter: • Nonlinear Operation • Replace outliers (impulsive noise, salt-pepper noise) with typical gray level value • Often preserves edge discontinuities (vs. mean filter) Other Order-Statistics Filters: • X percentile (vs. 50%) rank • Max filter (100%) • Min filter (0%)

© 2002 R. C. Gonzalez & R. E. Woods

Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Chapter 3: Intro. to Derivative Operations

Typical responses for 1st and 2nd derivative operators in 1-D

© 2002 R. C. Gonzalez & R. E. Woods

Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Chapter 3: Intro. to Derivative Operations - II

Discrete derivate operators for 1-D f  f ( x  1)  f ( x) x 2 f  f ( x  1)  f ( x)  f ( x)  f ( x  1) 2 x  f ( x  1)  2 f ( x)  f ( x  1)



© 2002 R. C. Gonzalez & R. E. Woods

 



Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Chapter 3: Intro. to Derivative Operations - III

Discrete derivate operators for 2-D f ( x, y )  f ( x  1, y )  f ( x, y ) x f ( x, y )  f ( x, y  1)  f ( x, y ) y  2 f ( x, y )  f ( x  1, y )  2 f ( x, y )  f ( x  1, y ) 2 x  2 f ( x, y )  f ( x, y  1)  2 f ( x, y )  f ( x, y  1) 2 y 2 2  f ( x , y )  f ( x, y ) 2  f   2 x y 2

© 2002 R. C. Gonzalez & R. E. Woods

Laplacian

Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Chapter 3: Re-Visit Derivative Operations

Typical responses for 1st and 2nd derivative operators in 1-D

© 2002 R. C. Gonzalez & R. E. Woods

Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Chapter 3: Use of Laplacian for Enhancement

+

90°

45°

 2 f  f ( x  1, y )  f ( x  1, y )  f ( x, y  1)  f ( x, y  1)  4 f ( x, y ) © 2002 R. C. Gonzalez & R. E. Woods

Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Chapter 3: Application of Laplacian for Enhancement

g ( x, y)  f ( x, y)  2 f Original plus high-pass filtered (Laplacian with negative center value in the mask)

g ( x, y)  5 f ( x, y)  f ( x  1, y)  f ( x  1, y)  f ( x, y  1)  f ( x, y  1) © 2002 R. C. Gonzalez & R. E. Woods

Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Chapter 3: Comparison of 2 Laplacian Masks for Enhancement with Combined Masks

4+1 8+1

© 2002 R. C. Gonzalez & R. E. Woods

Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Chapter 3: Unsharp Masking and High-Boost Filtering Unsharp Masking: Original – Blurred Version

f s ( x, y)  f ( x, y)  f ( x, y) High-Boost: A x Original – Blurred Version (A>1)

f hb ( x, y)  Af ( x, y)  f ( x, y) Using Laplacian: Blurred=Original – Sharpened Image

f hb ( x, y )  Af ( x, y )  f ( x, y )



 ( A  1) f ( x, y )  f ( x, y )  f ( x, y )



 ( A  1) f ( x, y )  f ( x, y )   2 f ( x, y )  Af ( x, y )   2 f ( x, y ) © 2002 R. C. Gonzalez & R. E. Woods

Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Chapter 3: Unsharp Masking and High-Boost Filtering Unsharp Masking: Original – Blurred Version

f s ( x, y)  f ( x, y)  f ( x, y) High-Boost Using Laplacian:

f hb ( x, y)  Af ( x, y)  2 f ( x, y)

© 2002 R. C. Gonzalez & R. E. Woods

Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Chapter 3: Example: Unsharp, High-Boost

© 2002 R. C. Gonzalez & R. E. Woods

Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Chapter 3: Use of First Derivatives for Enhancement: Gradient Vector, Gradient

f  2 1/ 2 2   f   f     y  Gx  f      f         f  G y    x   y     y   f  Gx  G y common approx .

© 2002 R. C. Gonzalez & R. E. Woods

Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Chapter 3: Masks for First Partial Derivatives

Roberts Operators

Sobel Operators

© 2002 R. C. Gonzalez & R. E. Woods

Digital Image Processing, 2nd ed.

www.imageprocessingbook.com

Chapter 3: Example use of Gradient

Use Gradient for preprocessing before automated inspection

© 2002 R. C. Gonzalez & R. E. Woods

Digital Image Processing, 2nd ed.

© 2002 R. C. Gonzalez & R. E. Woods

www.imageprocessingbook.com

Digital Image Processing, 2nd ed.

© 2002 R. C. Gonzalez & R. E. Woods

www.imageprocessingbook.com