Sigma Filter Research Articles

APPLICATION OF PALSAR-2 REMOTE SENSING DATA FOR LANDSLIDE HAZARD MAPPING IN KELANTAN RIVER BASIN, PENINSULAR MALAYSIA

Abstract. Yearly, several landslides ensued during heavy monsoons rainfall in Kelantan river basin, peninsular Malaysia, which are obviously connected to geological structures and topographical features of the region. In this study, the recently launched Phased Array type L-band Synthetic Aperture Radar-2 (PALSAR-2) onboard the Advanced Land Observing Satellite-2 (ALOS-2), remote sensing data were used to map geological structural and topographical features in the Kelantan river basin for identification of high potential risk and susceptible zones for landslides. Adaptive Local Sigma filter was selected and applied to accomplish speckle reduction and preserving both edges and features in PALSAR-2 fine mode observation images. Different polarization images were integrated to enhance geological structures. Additionally, directional filters were applied to the PALSAR-2 Local Sigma resultant image for edge enhancement and detailed identification of linear features. Several faults, drainage patterns and lithological contact layers were identified at regional scale. In order to assess the results, fieldwork and GPS survey were conducted in the landslide affected zones in the Kelantan river basin. Results demonstrate the most of the landslides were associated with N-S, NNW-SSE and NE-SW trending faults, angulated drainage pattern and metamorphic and Quaternary units. Consequently, structural and topographical geology maps were produced for Kelantan river basin using PALSAR-2 data, which could be broadly applicable for landslide hazard mapping.

Application of PALSAR-2 remote sensing data for structural geology and topographic mapping in Kelantan river basin, Malaysia

Natural hazards of geological origin are one of major problem during heavy monsoons rainfall in Kelantan state, peninsular Malaysia. Several landslides occur in this region are obviously connected to geological and topographical features, every year. Satellite synthetic aperture radar (SAR) data are particularly applicable for detection of geological structural and topographical features in tropical conditions. In this study, Phased Array type L-band Synthetic Aperture Radar (PALSAR-2), remote sensing data were used to identify high potential risk and susceptible zones for landslide in the Kelantan river basin. Adaptive Local Sigma filter was selected and applied to accomplish speckle reduction and preserving both edges and features in PALSAR-2 fine mode observation images. Different polarization images were integrated to enhance geological structures. Additionally, directional filters were applied to the PALSAR-2 Local Sigma resultant image for edge enhancement and detailed identification of linear features. Several faults, drainage patterns and lithological contact layers were identified at regional scale. In order to assess the results, fieldwork and GPS survey were conducted in the landslide affected zones in the Kelantan river basin. Results demonstrate the most of the landslides were associated with N-S, NNW-SSE and NE-SW trending faults, angulate drainage pattern and metamorphic and Quaternary units. Consequently, geologic structural map were produced for Kelantan river basin using recent PALSAR-2 data, which could be broadly applicable for landslide hazard assessment and delineation of high potential risk and susceptible areas. Landslide mitigation programmes could be conducted in the landslide recurrence regions for reducing catastrophes leading to economic losses and death.

SNR Improvement by Photon Noise Filtering in Ocean Color Monitor Satellite Images

In high radiometric resolution electro optical image payloads of remote sensing satellites, photon noise dominates SNR performance. Photon noise is input signal dependent and difficult to filter. This paper proposes a photon noise filtering technique for Ocean Color Monitor (OCM) images. Existing filtering techniques are meant for object detection and handles images with poor SNR. As OCM SNR is on higher side, custom sigma filter based denoising technique is developed. Proposed technique first converts photon noise to signal independent Gaussian noise. For this variance stabilization, Anscombe transform is used. Simulations are carried on various images. Proposed technique provides 2050% reduction in overall as well count-wise RMSE. FFT analysis shows significant reduction in noise. Proposed technique is of low complexity.

Polarimetric SAR Speckle Filtering and the Extended Sigma Filter

The advancement of synthetic aperture radar (SAR) technology with high-resolution and quad-polarization data demands better and efficient polarimetric SAR (PolSAR) speckle-filtering algorithms. Two requirements on PolSAR speckle filtering are proposed: 1) speckle filtering should be applied to distributed media only, and strong hard targets should be kept unfiltered; and 2) scattering mechanism preservation should be taken into consideration, in addition to speckle reduction. The purpose of this paper is twofold: 1) to propose an effective algorithm that is an extension of the improved sigma filter developed for single-polarization SAR; and 2) to investigate speckle characteristics and the need for speckle filtering for very high resolution (decimeter) PolSAR data. The proposed filter was specifically developed to account for the aforementioned two requirements. Its effectiveness is demonstrated with Jet Propulsion Laboratory airborne synthetic aperture radar data, and comparisons are made with a boxcar filter, the refined Lee filter, and a Wishart-based nonlocal filter. For very high resolution PolSAR systems, such as the German Aerospace Center F-SAR and Japanese Pi-SAR2, with decimeter spatial resolution, we found that the complex Wishart distribution is still valid to describe PolSAR speckle characteristics of distributed media and that speckle filtering may be needed depending on the size of objects to be analyzed. F-SAR X-band data with 25-cm resolution is used for illustration.

A method to calculate zero-signature satellite laser ranging normal points for millimeter geodesy - a case study with Ajisai

High repetition-rate satellite laser ranging (SLR) offers new possibilities for the post-processing of the range measurements. We analyze 11 years of kHz SLR passes of the geodetic satellite Ajisai delivered by Graz SLR station (Austria) in order to improve the accuracy and precision of the principal SLR data product - normal points. The normal points are calculated by three different methods: 1) the range residuals accepted by the standard 2.5 sigma filter, 2) the range residuals accepted by the leading edge filter and 3) the range residuals given by the single corner cube reflector (CCR) panels of Ajisai. A comparison of the statistical parameters of the obtained results shows that the selection of the range measurements from the leading edge of the SLR data distribution allows to minimize the satellite signature effect and to reduce the average single-shot RMS per normal point from 15.44 to 4.85 mm. The optical distance between the leading edge mean reflection point and the satellite’s center of mass is 1,023 mm, RMS = 1.7 mm. Further, in addition, we utilize the complete attitude model of Ajisai during the post-processing which enables selection of the range measurements to the single CCR panels of the satellite and the formation of the normal points which most closely approximate the physical distance between the ground station and the center of mass of Ajisai. This method eliminates the satellite signature effect from the distribution of the post-fit range residuals and further improves the average single-shot RMS per normal point to 3.05 mm. The normal point RMS per pass is reduced from 2.97 to 0.06 mm - a value expected for a zero-signature satellite.

Efficient scene‐based method for real‐time non‐uniformity correction of infrared video sequences

A method combing simple linear and nonlinear filters is proposed for real-time ‘non-uniformity correction’ of infrared video sequences, which suppresses ‘ghosting’ artefacts due to both lack of motion and strong edges. In this ‘least mean square’ (LMS)-based method, a mean filter is used first, when the ‘fixed pattern noise’ (FPN) level is high, taking advantage of its noise smoothing capability. When the FPN level drops to a low level, a sigma filter is used instead to reduce edge smearing. The sigma filter is also used to detect abnormal pixels like dead pixels and pixels contaminated by impulse noise, in addition to adaptive adjustment of the learning rate, with no extra cost. Experiments with simulated data and real infrared sequences show that the proposed method outperforms several other LMS methods. It is of the same computational complexity as Scribner's method, which makes it a good candidate for real-time hardware implementation.

Refined Filtering of Interferometric Phase From InSAR Data

Radar interferometry has been widely applied in measuring the terrain height and its changes. The information about surface can be derived from phase interferograms. However, the inherent phase noise reduces the accuracy and reliability of that information. Hence, the minimization of phase noise is essential prior to the retrieval of surface information that is embedded in an interferometric phase. This paper presents a refined filter based on the Lee adaptive complex filter and the improved sigma filter that was originally developed for amplitude image filtering. The basic idea is to adaptively filter the interferometric phase according to the local noise level to minimize the loss of signal for a particular pattern of fringes, including such extreme cases as involving broken fringes, following the removal of undesired pixels. Ultimately, the goals are to preserve the fringe pattern, to reduce phase bias and deviation, to reduce the number of residues, and to minimize the phase error. The preservation of the fringe pattern is particularly of concern in areas of high frequency of fringe and large phase gradient corresponding to steep terrains. The proposed refined filter was validated using both simulated data and real interferometric data. Results demonstrate that the filtering performance is better than that of commonly used filters.

영상에 포함된 특징의 방향성을 적용한 시그마 필터의 잡음제거

In the realization of obtained image by various visual equipments, the addition of noise to the original image is a common phenomenon and the occurrence of the noise is practically impossible to prevent completely. Thus, the noise detection and reduction is an important foundational purpose. In this study, we detect the orientation about feature of images and estimate the level of noise variance based on the measurement of the relative proportion of the noise. Also, we apply the estimated level of noise to the sigma filter on noise reduction algorithm. And using the orientation about feature of images by weighted value, we propose the effective algorithm to eliminate noise. As a result, the proposed statistical noise reduction methodology provides significantly improved results over the usual sigma filtering and regardless of the estimated level of the noise variance.

Approximation of weighted local mean operators

Recently, weighted local mean operators are widely used in image processing, compressive sensing and other areas. A weighted local mean operator changes its characteristics depending on a function content within a local area in order to preserve the function features. The directional diffusion filter and Yaroslavsky neighbourhood filter (also called the sigma filter) are discrete versions of such operators. Although these operators are not convolution ones, due to their sparsity, the corresponding numerical algorithms have simple structure and fast performance. In this paper, we study the approximate properties of the weighted local mean operators, particularly focus on their asymptotic expansions, which are related to non-linear diffusion equations.

Efficient noise reduction in images using directional modified sigma filter

A noise reduction of images using a directional modified sigma filter is proposed. It is important that an image should include accurate values without noise for large-scale data processing of a cloud computing environment. A conventional sigma filter has been shown to be a good solution both in terms of filtering accuracy and computational complexity. However, the sigma filter does not preserve small edges well especially for the high level of additive noise. In this paper, we propose a new method using a modified sigma filter. In our proposed method, an input image is first decomposed into two components that have features of horizontal, vertical, and diagonal direction. Then two components are applied: high-pass filtering (HPF) and low-pass filtering (LPF). By applying the conventional sigma filter separately on each of them, an output image is reconstructed from the filtered components. Added noise is removed and our proposed method preserves the edges in the image. Comparative results from experiments show that the proposed algorithm achieves higher gains than the sigma filter and modified sigma filter, which are 2.6 dB PSNR on average and 0.5 dB PSNR, respectively. When relatively high levels of noise are added, the proposed algorithm shows better performance than the two conventional filters. The proposed method can be efficiently applied in digital cameras, digital TV, and smart phones.

Shadow Detecting and Shadow Interpolation Algorithm for InSAS

The shadow detecting algorithm based on the coherence and the Sigma filter is used to pick up the shadow of interferometric synthetic aperture sonar (InSAS), which can eliminate small separated shadow areas. To solve the problems such as great computer complexity of traditional Shepard interpolation method and large fluctuant of linear interpolation method for the large shadow area, an improved Shepard interpolation method is proposed. Interpolation boundary is picked up by using diffuse search, and interpolation source is adaptively chosen according to the size of shadow area. The method carries out a perfect tradeoff between performance and computer speed. Lake trial dataset is used to validate the performance of proposed method. The results indicate that the proposed method can eliminate the fluctuant from the linear interpolation method and can process in real time in the InSAS system.

Refined sigma filter for SAR image despeckling using reference image

Presented is a refined sigma filter for synthetic aperture radar (SAR) image despeckling, where a novel reference image is constructed to replace the original SAR image for the selection of homogeneous pixels. The reference image holds the advantages of containing almost the same amount of details as the original SAR image but with less speckle noise, and thus enables more accurate sigma selection. Results on synthetic and real SAR images show the effectiveness of the method on both detail preservation and speckle removal.

Speckles Suppression Techniques for Ultrasound Images

Most medical images have a poorer signal to noise ratio than scenes taken with a digital camera, which often leads to incorrect diagnosis. Speckles suppression from ultrasound images is one of the most important concerns in computer-aided diagnosis. This article proposes two novel, robust and efficient ultrasound images denoising techniques. The first technique is the enhanced ultrasound images denoising (EUID) technique, which estimates automatically the speckle noise amount in the ultrasound images by estimating important input parameters of the filter and then denoising the image using the sigma filter. The second technique is the ultrasound image denoising using neural network (UIDNN) that is based on the second-order difference of pixels with adaptive threshold value in order to identify random valued speckles from images to achieve high efficient image restoration. The performances of the proposed techniques are analyzed and compared with those of other image denoising techniques. The experimental results show that the proposed techniques are valuable tools for speckles suppression, being accurate, less tedious, and preventing typical human errors associated with manual tasks in addition to preserving the edges from the image. The EUID algorithm has nearly the same peak signal to noise ratio (PSNR) as Frost and speckle-reducing anisotropic diffusion 1, whereas it achieves higher gains, on average—0.4 dB higher PSNR—than the Lee, Kuan, and anisotropic diffusion filters. The UIDNN technique outperforms all the other techniques since it can determine the noisy pixels and perform filtering for these pixels only. Generally, when relatively high levels of noise are added, the proposed algorithms show better performances than the other conventional filters.

Enhanced ultrasound images denoising technique

In ultrasound images, the noise can restrain information which is valuable for the general practitioner. Although sigma filter has been shown to be a good solution both in terms of filtering accuracy and computational complexity, it depends on estimating both the mean and variance manually. This paper presents an Enhanced Ultrasound Images Denoising (EUID) technique for speckle noise suppression in ultrasound images. This technique estimates automatically speckle noise amount in the ultrasound images by estimating important input parameters of the filter and then denoise the image using the sigma filter. The experimental results show that the proposed technique is a valuable tool for estimating speckle standard deviation, being accurate, less tedious and preventing typical human errors associated with manual tasks in addition to preserving the edges from the image. A comparative study is made on the performances of the EUID technique and other filters in removing the speckles from the image and in preserv...

Semi-automated procedures for shoreline extraction using single RADARSAT-1 SAR image

Coastline identification is important for surveying and mapping reasons. Coastline serves as the basic point of reference and is used on nautical charts for navigation purposes. Its delineation has become crucial and more important in the wake of the many recent earthquakes and tsunamis resulting in complete change and redraw of some shorelines. In a tropical country like Malaysia, presence of cloud cover hinders the application of optical remote sensing data. In this study a semi-automated technique and procedures are presented for shoreline delineation from RADARSAT-1 image. A scene of RADARSAT-1 satellite image was processed using enhanced filtering technique to identify and extract the shoreline coast of Kuala Terengganu, Malaysia. RADSARSAT image has many advantages over the optical data because of its ability to penetrate cloud cover and its night sensing capabilities. At first, speckles were removed from the image by using Lee sigma filter which was used to reduce random noise and to enhance the image and discriminate the boundary between land and water. The results showed an accurate and improved extraction and delineation of the entire coastline of Kuala Terrenganu. The study demonstrated the reliability of the image averaging filter in reducing random noise over the sea surface especially near the shoreline. It enhanced land-water boundary differentiation, enabling better delineation of the shoreline. Overall, the developed techniques showed the potential of radar imagery for accurate shoreline mapping and will be useful for monitoring shoreline changes during high and low tides as well as shoreline erosion in a tropical country like Malaysia.

Pre-processing of TerraSAR-X Data for Speckle Removal: An Approach for Performance Evaluation

Measures to select despeckling filter to remove speckles from TerraSAR-X image have been evaluated for an area having varied land cover classes. Gamma, Lee, Sigma, Mean, Median and Frost filters were applied in different window sizes. These images are visually compared and four parameters viz. Noise Mean Value (NMV), Noise Variance (NV), Mean Square Difference (MSD) and Equivalent Numbers of Looks (ENL) are calculated. These parameters were well implemented to compare results of various despeckling filters on TerraSAR-X image recorded in high resolution spot mode and checked for the constraints. Median 7x7 filter was found most suitable among other filters to despeckle the images.

Support vector machine approach to identifying buildings using multi-temporal ALOS/PALSAR data

Accurate detection of built-up areas is valuable for quantifying the level of urbanization and monitoring the decreasing amount of agricultural land. In this study, three-temporal, dual-polarization Advanced Land Observing Satellite (ALOS)/Phased Array type L-band Synthetic Aperture Radar (PALSAR) images were integrated to map buildings in the Yangtze River Delta of East China, where land has been intensively used. The results show that the support vector machine (SVM) classifier performs well in identifying buildings, with an accuracy of 90% in urban areas and 95% in rural areas, even with only a small number of training samples. Buildings in urban areas are more likely to be underestimated (commission error of 15%) than those in a rural environment. Visual inspection and quantitative analysis confirmed that the Local Sigma Filter considerably reduced random speckle noise in the PALSAR imagery. Thus, the filter is suitable for enhancing feature extraction of future multi-polarization and multi-temporal SAR imagery. Overall, the buildings identification approach proposed in this study could serve as a valuable tool for operational monitoring of rural land use change and urban sprawl.

SAR Image Despeckling Using Bayesian Nonlocal Means Filter With Sigma Preselection

Bayesian nonlocal (NL) (BNL) means filter, as an extension of the NL means algorithm, provides a general framework for image denoising when dealing with different noise. However, this approach makes a strong assumption that image patch itself provides a good approximation on the true parameter, which leads to the bias problem particularly under serious speckle noise. Another disadvantage of the BNL filter is that the commonly used patch preselection method cannot effectively exclude the outliers. In this letter, a new form of the BNL filter is presented for the purpose of synthetic aperture radar image despeckling, which incorporates the technique of sigma filter to cope with the bias problem. In addition, pixel preselection is adopted based on the refined sigma range, which greatly contributes to the preservation of the image details such as edges, texture, and the strong reflective scatters. Experimental results illustrate that the proposed BNL filter reaches the state-of-the-art performance on both the visual quality and evaluation indexes.

Construction of local nonlinear filter without staircase effect in image restoration

This article introduces an innovative approach to the construction of a local nonlinear filter in image processing, which eliminates staircase effect. A local nonlinear filter changes its characteristics depending on the image content within a local area. Hence, it may preserve the image features. Due to its local property, the filtering algorithm has simple structure and fast performance. Two well-known local nonlinear filters are Yaroslavsky neighbourhood filter (also called sigma filter) and the bilateral filter (also called SUSAN filter). These filters produce very fast and effective denoising algorithms. But, unfortunately, they show staircase effect. To overcome their drawback, in this article we construct a new local nonlinear filter, called directional diffusion filter (DDF), which preserves image features and does not show staircase effect. We also study the DDF's properties and reveal the relation between DDF and Rudin–Osher–Fatemi's total variation model. Since the DDF denoising algorithm is very fast and uses very little read-only memory (ROM), it can be applied in the real-time processing for the devices without plenty of ROM, such as cell phones, security cameras and multispectral imagery sensors.

Objective and Subjective Evaluation of Spatially Transcoded Videos for Mobile Receivers

Digital television produces video signals with different bit rates, encoding formats, and spatial resolutions. To deliver video to users with different receivers, the content needs to be dynamically adapted. Transcoding devices convert video from one format into another. The reception of digital videos using mobile receivers, implies that the spatial resolution of the video must be adjusted to fit the small display. This paper presents subjective and objective quality analysis of spatially transcoded videos. Transcoding algorithms that downsample the video frames using the moving average, median, mode, weighted average and sigma filters are considered.