![Indonesian Journal of Electrical Engineering and Computer Science
Vol. 8, No. 3, December 2017, pp. 654 ~ 656
DOI: 10.11591/ijeecs.v8.i3.pp654-656 654
Received August 9, 2017; Revised October 20, 2017; Accepted November 18, 2017
GIS Based Satellite Image Denoising Using Curvelet
Transform
S. Sinthuja, S. Saravanan
Department of Electrical and Electronics Engineering ( Marine), AMET University, Chennai, India
Abstract
Generally, satellite images contain very significant information about geographical features such
as rivers, roads, building and bridges etc of the earth. Geographic Information System (GIS) requires these
features for automatic detection and it has been corrupted by various types of noise. Curvelet Transform
(CT) is used in the proposed system for denoising the images. Advantages of multi resolution image such
as line, compatibility of human visual system and edge detection are provided. Then K-Means clustering is
used in this system for segmentation purpose after the pre processing done.
First, K-Means algorithm is used for segmenting background and water then extraction of bridges
is done based on pixel intensity difference.
Keywords: Satellite Image denoising, CT, K-Means, PSNR
Copyright © 2017 Institute of Advanced Engineering and Science. All rights reserved.
1. Introduction
Shearlet transform based satellite image denoising is presented in [1]. Shearlet
transform is used for denoising the satellite images and Artificial Bee Colony (ABC) optimization
is used for selecting the threshold value for denoising. Linear Discriminant Analysis (LDA)
based on clustering for SAR Image denoising is described in [2]. Noise clustering is the main
task for denoising SAR images into various disjoint local regions and again it denoised every
splitted regions using wiener filtering in domain LDA. Dual Tree Complex Wavelet Transform
(DT-CWT) based satellite image enhancement for denoising and illumination enhancement is
discussed in [3]. Noisy input images are decomposed into various frequency sub bands using
DT-CWT and these sub bands are denoised using local adaptive bivariate shrinkage function.
DT-CWT is used to decompose the denoised image again into different frequency sub bands.
Finally, inverse DT-CWT reconstructs the image.
Satellite image clustering in HSV and RGB Color Space based on k-means is explained
in [4]. Here RGB color space performance is decreased by using k-means when compared to
HSV color space. More number of color spaces is compared in this method. Evolutionary
algorithms based enhanced sub-band adaptive thresholding function for satellite image
denoising is described in [5]. ABC, Cuckoo Search (CS) and particle swarm optimization is the
stochastic global optimization techniques for parameters learning of adaptive thresholding
function. ABC and CS approach are used for denoising. Weiner filter with SPEA2 algorithm
based satellite image denoising is presented in [6]. Land use and land cover classification of
LISS-III satellite image using KNN and decision tree is presented in [7]. High noise level and pre
filtering are removed by using Weiner filter with SPEA2 algorithm and it denoised the images
then improved algorithm is extended. Finally PSNR and SNR are calculated. Genetic algorithm
based image super resolution using wavelet transform is implemented in [8].
2. Methodology
2.1. Curvelet Transform
CT is used in the proposed work for denoising the satellite images for detection of
bridges. It is a multi scale multidirectional transform which can consistently rebuild line shaped
edge features. More complex curvelet is the first generation curvelet due to the need of Ridgelet
transform. Spatially selected Fourier samples and wrapping based computation is the second](https://image.slidesharecdn.com/169252eeem-gisbasedsatelliteimagedenoisingeditari-201016080550/85/16-9252-eeem-gis-based-satellite-image-denoising-edit-ari-1-320.jpg)
![ ISSN: 2502-4752
IJEECS Vol. 8, No. 3, December 2017 : 654 – 656
656
References
[1] Anju TS, Raj NN. Satellite Image Denoising Using Shearlet Transform. IEEE International Conference
on Communication and Signal Processing. 2016: 0571-0575.
[2] Rajapriyadharshini R. SAR Image Denoising via Clustering Based Linear Discriminant Analysis. IEEE
International Conference on Innovations in Information, Embedded and Communication Systems.
2015: 1-6.
[3] Taşmaz H, et al. Satellite Image Enhancement by Using Dual Tree Complex Wavelet Transform:
Denoising and Illumination Enhancement. IEEE 20th Signal Processing and Communications
Applications Conference. 2012: 1-4.
[4] Kumar G. et al. Performance of K-means Based Satellite Image Clustering in RGB and HSV Color
Space. IEEE International Conference on Recent Trends in Information Technology. 2016: 1-5.
[5] Soni V. et al. Improved sub-band Adaptive Thresholding Function for Denoising of Satellite Image
Based on Evolutionary Algorithms. IET Signal Processing. 2013; 7(8): 720-730.
[6] Khajwaniya KK, Tiwari V. Satellite Image Denoising Using Weiner Filter with SPEA2 Algorithm. IEEE
9th International Conference on Intelligent Systems and Control. 2015: 1-6.
[7] Upadhyay A, Shetty A, Singh SK, Siddiqui Z. Land Use and Land Cover Classification of LISS-III
Satellite Image Using KNN and Decision Tree. In Computing for Sustainable Global Development
(INDIACom), 2016 3rd International Conference on IEEE, March 2016: 1277-1280.
[8] Panda SS, Jena G. Image Super Resolution Using Wavelet Transformation Based Genetic Algorithm.
In Computational Intelligence in Data Mining. Springer India. 2016; 2: 355-361.](https://image.slidesharecdn.com/169252eeem-gisbasedsatelliteimagedenoisingeditari-201016080550/85/16-9252-eeem-gis-based-satellite-image-denoising-edit-ari-3-320.jpg)
16 9252 eeem gis based satellite image denoising (edit ari)
- 1. Indonesian Journal of Electrical Engineering and Computer Science Vol. 8, No. 3, December 2017, pp. 654 ~ 656 DOI: 10.11591/ijeecs.v8.i3.pp654-656 654 Received August 9, 2017; Revised October 20, 2017; Accepted November 18, 2017 GIS Based Satellite Image Denoising Using Curvelet Transform S. Sinthuja, S. Saravanan Department of Electrical and Electronics Engineering ( Marine), AMET University, Chennai, India Abstract Generally, satellite images contain very significant information about geographical features such as rivers, roads, building and bridges etc of the earth. Geographic Information System (GIS) requires these features for automatic detection and it has been corrupted by various types of noise. Curvelet Transform (CT) is used in the proposed system for denoising the images. Advantages of multi resolution image such as line, compatibility of human visual system and edge detection are provided. Then K-Means clustering is used in this system for segmentation purpose after the pre processing done. First, K-Means algorithm is used for segmenting background and water then extraction of bridges is done based on pixel intensity difference. Keywords: Satellite Image denoising, CT, K-Means, PSNR Copyright © 2017 Institute of Advanced Engineering and Science. All rights reserved. 1. Introduction Shearlet transform based satellite image denoising is presented in [1]. Shearlet transform is used for denoising the satellite images and Artificial Bee Colony (ABC) optimization is used for selecting the threshold value for denoising. Linear Discriminant Analysis (LDA) based on clustering for SAR Image denoising is described in [2]. Noise clustering is the main task for denoising SAR images into various disjoint local regions and again it denoised every splitted regions using wiener filtering in domain LDA. Dual Tree Complex Wavelet Transform (DT-CWT) based satellite image enhancement for denoising and illumination enhancement is discussed in [3]. Noisy input images are decomposed into various frequency sub bands using DT-CWT and these sub bands are denoised using local adaptive bivariate shrinkage function. DT-CWT is used to decompose the denoised image again into different frequency sub bands. Finally, inverse DT-CWT reconstructs the image. Satellite image clustering in HSV and RGB Color Space based on k-means is explained in [4]. Here RGB color space performance is decreased by using k-means when compared to HSV color space. More number of color spaces is compared in this method. Evolutionary algorithms based enhanced sub-band adaptive thresholding function for satellite image denoising is described in [5]. ABC, Cuckoo Search (CS) and particle swarm optimization is the stochastic global optimization techniques for parameters learning of adaptive thresholding function. ABC and CS approach are used for denoising. Weiner filter with SPEA2 algorithm based satellite image denoising is presented in [6]. Land use and land cover classification of LISS-III satellite image using KNN and decision tree is presented in [7]. High noise level and pre filtering are removed by using Weiner filter with SPEA2 algorithm and it denoised the images then improved algorithm is extended. Finally PSNR and SNR are calculated. Genetic algorithm based image super resolution using wavelet transform is implemented in [8]. 2. Methodology 2.1. Curvelet Transform CT is used in the proposed work for denoising the satellite images for detection of bridges. It is a multi scale multidirectional transform which can consistently rebuild line shaped edge features. More complex curvelet is the first generation curvelet due to the need of Ridgelet transform. Spatially selected Fourier samples and wrapping based computation is the second
- 2. IJEECS ISSN: 2502-4752 GIS Based Satellite Image Denoising Using Curvelet Transform (S. Sinthuja) 655 generation. Curvelet characteristics are a good one used to denoising the satellite images is proposed. 2.2. Bridge Detection There are different steps are involved in this bridge detection algorithm for segmentation. First step of detection is the segmentation using k-means used to differentiate the background and foreground region. Then histogram techniques also used to get more accurate results based on k-means. Block diagram of the proposed satellite image denoising is shown in figure below. Figure 1. Block diagram of the proposed satellite image denoising technique 3. Experimental Results SAR image is given as input shown in the figure. It is generally suffer from the noise. CT is proposed in this paper to denoise the image. This section tells about proposed denoising scheme using CT is exploited for satellite images. Performance of the proposed scheme is computed by PSNR value. Figure 2 shows the proposed (a) noisy image and (b) denoised image. (a) (b) Figure 2. (a) Noisy image (b) Denoised image 4. Conclusion A novel approach to denoise the satellite images was developed using thresholding based on CT is proposed. From the investigational results the image denoised using the proposed method was found to be more visually tempting than other existing algorithms. Figure 2 shows the proposed bridge detection and denoised image using k-means technique. The results show that k-means method is a very efficient optimization and obtained PSNR value is 32.52.
- 3. ISSN: 2502-4752 IJEECS Vol. 8, No. 3, December 2017 : 654 – 656 656 References [1] Anju TS, Raj NN. Satellite Image Denoising Using Shearlet Transform. IEEE International Conference on Communication and Signal Processing. 2016: 0571-0575. [2] Rajapriyadharshini R. SAR Image Denoising via Clustering Based Linear Discriminant Analysis. IEEE International Conference on Innovations in Information, Embedded and Communication Systems. 2015: 1-6. [3] Taşmaz H, et al. Satellite Image Enhancement by Using Dual Tree Complex Wavelet Transform: Denoising and Illumination Enhancement. IEEE 20th Signal Processing and Communications Applications Conference. 2012: 1-4. [4] Kumar G. et al. Performance of K-means Based Satellite Image Clustering in RGB and HSV Color Space. IEEE International Conference on Recent Trends in Information Technology. 2016: 1-5. [5] Soni V. et al. Improved sub-band Adaptive Thresholding Function for Denoising of Satellite Image Based on Evolutionary Algorithms. IET Signal Processing. 2013; 7(8): 720-730. [6] Khajwaniya KK, Tiwari V. Satellite Image Denoising Using Weiner Filter with SPEA2 Algorithm. IEEE 9th International Conference on Intelligent Systems and Control. 2015: 1-6. [7] Upadhyay A, Shetty A, Singh SK, Siddiqui Z. Land Use and Land Cover Classification of LISS-III Satellite Image Using KNN and Decision Tree. In Computing for Sustainable Global Development (INDIACom), 2016 3rd International Conference on IEEE, March 2016: 1277-1280. [8] Panda SS, Jena G. Image Super Resolution Using Wavelet Transformation Based Genetic Algorithm. In Computational Intelligence in Data Mining. Springer India. 2016; 2: 355-361.