edge detection of image

Views:
 
Category: Entertainment
     
 

Presentation Description

edge detection by various method

Comments

Presentation Transcript

A Comparative Study on Image Edge Enhancement for Synthetic Aperture Radar (SAR) Images :

A Comparative Study on Image Edge Enhancement for Synthetic Aperture Radar (SAR) Images Presented By Sk Sariful Islam Ahammed M-Tech, 2 nd Year Roll no.-001 Dept of Computer Sc. & Engg. Narula Institute of Technology

Outline:

Outline Aim of the work Property of SAR image Proposed novel edge enhancement algorithm based on 2D wavelet transform The sobel edge detected image is improved via a fuzzy-genetic optimization Experimental results & discussion Conclusions

Aim of the work:

Aim of the work Develop a novel two novel methods for edge enhancement in SAR images, of which one is based on the exploitation of the information provided by the wavelet coefficients and another is based on a hybrid fuzzy-genetic optimization approach Then Remove speckle noise

Property of Synthetic Aperture Radar (SAR) Images :

Property of Synthetic Aperture Radar (SAR) Images Speckle Noise: 1. Due to out of phase 2. Type of Multiplicative noise Irregularity and Discontinuity 1.Due to Speckle noise

First Approach :

First Approach

Image Acquisition :

Image Acquisition SAR image is taken as input for processing

PowerPoint Presentation:

Logarithm of image Here the input image is converted into gray scale image at first Then logarithmic transformation is induced

PowerPoint Presentation:

2D Wavelet Transformation Take SWT2 Wavelet Transform Haar Wavelet Four sub-bands Fig 3: Horizontal Sub Band Fig 4: Vertical Sub Band Fig 5: Diagonal Sub Band

Why SWT2:

Why SWT2 DWT is not time & translation invariant Extended version of DWT 1. average some slightly different DWT Better De-noising Edge Preserving Break Down Point Detection

Why Haar Wavelet:

Why Haar Wavelet Advantage: 1. Less time complexity Disadvantage: 1. not continuous

Normalization, Point wise Maximum and Multiplication:

Normalization, Point wise Maximum and Multiplication Normalize each sub-band to their maximum and take their absolute value 1. Repetition of normalization method ensures low values on the sub-bands Point-wise maximum means taking the maximum value pixel per pixel among matrices Create a single image or matrix by using point-wise maximum operation.The point-wise maximum operation is carried out with different iteration The different intermediate maxima are calculated which are combined through point wise multiplication

PowerPoint Presentation:

Edge Enhancement using First Approach

PowerPoint Presentation:

Second Approach Edge Enhancement via Second Method In this Approach Four methods have been used Sobel edge detection Fuzzy Logic based enhancement of edge image Optimization of edge using genetic algorithm Relaxed median filter edge detection

PowerPoint Presentation:

Sobel Edge Detection It returns edges at those points where the gradient of the input image has maximum

PowerPoint Presentation:

Fuzzy Logic based Enhancement After sobel edge detection technique ,we apply these one for edge detection. The whole test image has been scanned using a 2*2 pixel window. The different rules which has been used for marking the pixel under consideration as Black, White or Edge has been furnished in Table-1.

PowerPoint Presentation:

Contd.... Table-1 Fuzzy Rule Matrix

PowerPoint Presentation:

Optimization of Edge using Genetic Algorithm An Image has been taken as an input matrix denoted by A Then 4 pixels forming 2x2 matrixes from the left most corner of the image matrix has been taken. The fitness value of each pixel using fitness function is calculated. Decimal value of each pixel is converted into binary value. At the time of crossover two parent’s binary string is divided and combined. After creating offspring each of string’s binary value is converted into decimal value. Each decimal value with its corresponding old value is compared. If new value is greater than the old value, the old pixel value with the new obtained value is replaced. In this way the whole image by considering each population as 2x2 matrixes is traversed.

PowerPoint Presentation:

Image obtained by GA

PowerPoint Presentation:

Relaxed Median Filter Edge Detection Sharp corner, Thin line, Edge preserving This filter reduce some speckle noise in SAR images. The output is given below.

PowerPoint Presentation:

Experimental results & Discussion

PowerPoint Presentation:

Two methods are proposed for edge detection technique for Synthetic Aperture Radar (SAR) Images. The comparative study of Root Mean Square Error (RMSE) suggests that its value has been less in case of 2D wavelet based image edge enhancement scheme than that of other approaches. Conclusion

References:

References J. W. Goodman, Speckle Phenomena in Optics: Theory and Applications. Englewood, CO: Roberts & Company, 2007. Goodman, Some fundamental properties of speckle, J Opt. Soc. Amer., vol. 66, pp. 1145-1150, Nov. 1976. MarivíTello Alonso, Carlos López-Martínez, Jordi J. Mallorquí, and Philippe Salembier, “Edge Enhancement Algorithm Based on the Wavelet Transform for Automatic Edge Detection in SAR Images”. IEEE Transactions on geoscience and Remote Sensing, pp.222-235, 2011. H. Arsenault and G. April, “Properties of speckle integrated with a finite aperture and logarithmically transformed,” J. Opt. Soc. Amer., vol.66, no. 11, pp. 1160–1163, Nov. 1976.

Project :

Project An Approach of image enhancement in frequency domain: A Comparative Study

PowerPoint Presentation:

THANK YOU

authorStream Live Help