IMAGE ENHANCEMENT - POINT PROCESSING

Transcription

1 1 IMAGE ENHANCEMENT - POINT PROCESSING KOM3212 Image Processing in Industrial Systems Some of the contents are adopted from R. C. Gonzalez, R. E. Woods, Digital Image Processing, 2nd edition, Prentice Hall, 2008 Digital Image Processing: An Algorithmic Introduction using Java, W Burger, Mark J. Burge, Springer Verlag, 2008

2 2 Today s Lecture Digital image representation Image enhancement Spatial domain Image Enhancement methods Point-based methods 1. Image Negative 2. Log transformation 3. Power-law 4. Contrast stretching 5. Gray-level slicing 6. Bit plane slicing 7. Histogram equalization 8. Averaging

3 3 Digital image representation Going Digital: What is projected onto the image plane of the camera is essentially a 2D time dependent, continuous distribution of light energy. A digital image & image detail Different images types

4 4 Digital image representation

5 5 Digital image representation

6 6 Digital image representation

7 7 Digital image representation Best file types for these general purposes: Properties For Unquestionable Best Quality Smallest File Size Maximum Compatibility (PC, Mac, Unix) Worst Choice Photographic Images Photos are continuous tones, 24-bit color or 8-bit Gray, no text, few lines and edges TIF LZW or PNG (lossless compression and no JPG artifacts) JPG with a higher Quality factor can be both small decent quality. TIF or JPG 256 color GIF is very limited color, and is a larger file than 24 -bit JPG Graphics, including Logos or Line art Graphics are often solid colors, with few colors, limited to 256 colors, with text or lines and sharp edges PNG or TIF LZW (lossless compression, without JPG artifacts) TIF LZW or GIF or PNG (graphics/logos without gradients normally permit indexed color of 2 to 16 colors for smallest file size) TIF or GIF JPG compression adds artifacts, smears text and lines and edges

8 8 IMAGE ENHANCEMENT SPATIAL DOMAIN IMAGE ENHANCEMENT METHODS Some of the contents are adopted from R. C. Gonzalez, R. E. Woods, Digital Image Processing, 2nd edition, Prentice Hall, 2008 Digital Image Processing: An Algorithmic Introduction using Java, W Burger, Mark J. Burge, Springer Verlag, 2008

9 9 What is Image Enhancement? Processing an image for visual interpretation is mostly referred as Image Enhancement Visual evaluation of image quality is a highly subjective process Thus the definition of a good image is not always possible Human evaluation vs machine evaluation Performance of the algorithms has no clear-cuts

10 10 Different approaches for Image Enhancement Spatial domain Point-based processing (involved with gray levels of each pixel) (this is the category we are dealing with right now) Mask-based processing (neighbor-based processing, involved with spatial filters related operations) Frequency domain Frequency domain filters

11 11 Point based processing The spatial domain refers to the aggregate of pixels composing an image f(0,0) Spatial Domain f(x, y) g(x, y) = T f x, y f(x, y) is the input image g(x, y) is the output image T is an operator on f(x, y) s = T(r) r : is the set of gray level of input image s : is the set of gray levels of output image Image f

12 12 Point based processing

13 13 Image Negatives s = T(r) = L 1 r

14 14 Log transformation (Dynamic range compression) s = T r = c log(1 + r )

15 15 Power-Law transformation s = T r = cr γ

16 16 Power-Law transformation

17 17 Power-Law transformation- Gamma correction

18 18 Power-Law transformation- Gamma correction

19 19 Contrast stretching s T r mr b

20 20 Gray-level slicing Highlighting an intensity range s = T r = 255 if A r B 0 otherwise

21 21 Bit-plane slicing Highlighting the contribution made by a specific bit. For pgm images, each pixel is represented by 8 bits. Each bit-plane is a binary image 8.Bit (msb) slice 8. bit of slicethe original image

22 22 8. bit slice 7. bit slice 6. bit slice 5. bit slice 4. bit slice 3. bit slice 2. bit slice 1. bit slice

23 23 Histogram Processing Gray-level histogram is a function showing, for each gray level, the number of pixels in the image that have that gray level. number of pixel which has intensity r k n k = h r k = 1 f x,y =r k p k = n k Normalized histogram (probability): N k 0, 255 for monochrome gray level images

24 24 Dark Image Low Cont Image Bright Image Hight Cont. Image

25 25 Histogram - Examples

26 26 Histogram - Examples

27 27 Histogram - Examples

28 28 Histogram equalization Transformation function p r (r) is the probability density function (pdf) The transformation function is the cumulative distribution function (CDF) T(r) is single-valued and monotonically increasing within range of r T(r) has the same range as r To make the pdf of the transformed image uniform, i.e. to make the histogram of the transformed image uniform s k = T r k = n j N j=0..k = p r (r j ) j=0..k

29 29 Histogram equalization - Example Gray Level Value Frequency (count) PDF / / / / / / /90

30 30 Histogram equalization - Example Calculate CDF according to gray levels Gray Level Value PDF CDF 0 10/90 = /90 = /90 = /90 = floor[ CDF * (Levels- 1) ] Then in this step you will multiply the CDF value with (Gray levels -1) Considering we have an 3 bpp image. Then number of levels we have are 8. And 1 subtracts 8 is 7. So we multiply CDF by 7. Here what we got after multiplying. 4 10/90 = /90 = /90 =0.111 ~1 7

31 31 Histogram equalization - Example Old Histogram New Histogram Gray Level Value New Gray Level Value Frequency if we map old values to our new values, then this is what we got

32 32 Histogram equalization Dark Image Low Cont Image Bright Image Hight Cont. Image

33 33 Local enhancement

34 34 Image Averaging original image f x, y + noisy image g i x, y Noise η i x, y i=1...m (M images are in the question) Different acquisition of the scene g i x, y = f x, y + η i x, y then M 1 M 1 M 1 g i x, y = f x, y + η i x, y i=0 i=0 i=0 g x, y = f x, y + η x, y If the noise is uncorrelated and has zero expectation, then the expectation of g x, y will be f x, y E [g x, y ] = f x, y