Vladimir Botchko [email protected]

16
1 Vladimir Botchko [email protected] Lecture 3. Image Lecture 3. Image Enhancement in Spatial Enhancement in Spatial Domain Domain Lappeenranta University of Technology (Finland)

description

Lappeenranta University of Technology (Finland). Lecture 3. Image Enhancement in Spatial Domain. Vladimir Botchko [email protected]. Image Enhancement. Simple intensity transformations Histogram processing (equalization) Image subtraction Image averaging - PowerPoint PPT Presentation

Transcript of Vladimir Botchko [email protected]

Page 1: Vladimir Botchko  botchko@lut.fi

1

Vladimir Botchko [email protected]

Lecture 3. Image Enhancement in Lecture 3. Image Enhancement in Spatial DomainSpatial Domain

Lappeenranta University of Technology (Finland)

Page 2: Vladimir Botchko  botchko@lut.fi

2

Image Enhancement

Simple intensity transformations Histogram processing (equalization) Image subtraction Image averaging Spatial filtering (smoothing, sharpening) Enhancement. First derivative (gradient)

Example of manual edge enhancement: made by artist.

Page 3: Vladimir Botchko  botchko@lut.fi

3

Simple intensity transformations

Contrast stretched

Page 4: Vladimir Botchko  botchko@lut.fi

4

Simple intensity transformations

Intensity level slicing: Original image (top) Thresholded (left) Gray-level slicing (lower right)

Page 5: Vladimir Botchko  botchko@lut.fi

5

Image Enhancement

Simple intensity transformations Histogram processing Image subtraction Image averaging Spatial filtering (smoothing, sharpening) Enhancement. First derivative (gradient).

Page 6: Vladimir Botchko  botchko@lut.fi

6

Histogram equalization

Page 7: Vladimir Botchko  botchko@lut.fi

7

Image Enhancement

Simple intensity transformations Histogram processing Image subtraction Image averaging Spatial filtering (smoothing, sharpening) Enhancement. First derivative (gradient).

Page 8: Vladimir Botchko  botchko@lut.fi

8

Image subtraction

Page 9: Vladimir Botchko  botchko@lut.fi

9

Image Enhancement

Simple intensity transformations Histogram processing Image subtraction Image averaging Spatial filtering (smoothing, sharpening) Enhancement. First derivative (gradient).

Page 10: Vladimir Botchko  botchko@lut.fi

10

Image averaging. Spatial filtering (smoothing)

Original image (upper left) Original + noise (upper right) Smoothed image (lower right) Median smoothing (lower left)

Page 11: Vladimir Botchko  botchko@lut.fi

11

Order-statistics filter for binary images (when numl=5 then it is

median) Rank filter is used for smoothing after recognition for segmentation. Simple Matlab

code is here:

Page 12: Vladimir Botchko  botchko@lut.fi

12

Order-statistics filter

First is median filtering result (rank 5 for 3x3 window). Upper part is input image lower part is smoothed image

Page 13: Vladimir Botchko  botchko@lut.fi

13

Order-statistics filter

Then the rank (rank 3 for 3x3 window). Original image upper, smoothed image is lower

Page 14: Vladimir Botchko  botchko@lut.fi

14

Sharpening

Second Derivatives. Laplacian

Page 15: Vladimir Botchko  botchko@lut.fi

15

Image Enhancement

Simple intensity transformations Histogram processing Image subtraction Image averaging Spatial filtering (smoothing, sharpening) Enhancement. First derivative (gradient).

Page 16: Vladimir Botchko  botchko@lut.fi

16

Median filter and Sobel operator

Gray-level images (from left to right from top to bottom): original texture

synthetic (computer generated) image, the result of recognition corrupted by classification errors, median filtering, edge extraction through Sobel operator, superposition a gray level image and image with edges.