Noise Elimination Method Research Articles

Contactless Continuous Activity Recognition based on Meta-Action Temporal Correlation in Mobile Environments

Continuous activity recognition (CAR) plays an important role in human daily indoor activity monitoring and can be widely used in smart home, human-computer interaction and user authentication. Due to the privacy issue and limited coverage of video signals, RF-based CAR has attracted more and more attention in recent years. This paper focuses on three key problems in RF-based CAR: denoising, segmentation and recognition. We present the design and implementation of a contactless and sensorless continuous activity recognition system, namely WiCheck. Our basic idea is to utilize the temporal correlation between two adjacent actions in continuous activity to eliminate the cumulative error in continuous activity segmentation. Firstly, the multi-layer optimized noise elimination method is used to decrease the environment interference. Secondly, a method based on dual-swing window is proposed to reduce the cumulative error of continuous activity segmentation. Finally, WiCheck is implemented in different indoor environments, and 6 continuous activity sequences are designed to evaluate and analyze the influencing factors. The continuous activity recognition accuracy of WiCheck to two actions and three actions can approach 90% and 75%, respectively.

노이즈 제거 및 전압 위상 지연을 고려한 부분방전 패턴 인식

In this study, the effect of noise elimination and voltage phase lag on the partial discharge pattern recognition was studied to improve the accuracy of the partial discharge diagnosis of the XLPE transmission cable system. Recognition rates were compared by applying Neural Network and SVM techniques using statistical feature values extracted from physical quantities in PRPDA data such as discharge numbers and discharge amounts according to the voltage phase, which were measured through the commercial partial discharge diagnostic system for four different types of partial discharge models. As a result, it was found that the minimum noise elimination method showed high pattern recognition rate because it relatively preserved the partial discharge information, even though the background noise would not be clearly eliminated. In addition, for the effect of the voltage phase lag, the neural network did not show any meaningful effect, whereas SVM showed significantly lowered recognition rate.

Regression-Based Neuro-Fuzzy Network Trained by ABC Algorithm for High-Density Impulse Noise Elimination

Salt and pepper (SAP) noise elimination is a crucial step for further image processing and pattern recognition applications. The main aim of this article is to propose a novel SAP noise elimination method which employs a regression-based neuro-fuzzy network for highly corrupted gray scale and color images. In the proposed method, multiple neuro-fuzzy filters trained with artificial bee colony algorithm is combined with a decision tree algorithm. The performance of the proposed filter is compared with a number of well known methods with respect to popular metrics including, structural similarity index, peak signal-to-noise ratio, and correlation on well known test images. The results reveal that the proposed filter has superior performance in terms of all comparison metrics.

Blind cyclic feature detection with noise elimination in cognitive radios with fast fading

A practical implementation and the maximum likelihood detection analysis of the blind cyclic-feature detection (BCD) in cognitive radios is presented in this study. The sum of the magnitude squared of the cyclic spectral density (SCSD) is the merit criterion for the maximum likelihood detection. The authors show that the SCSD can be conveniently obtained from the autocorrelation of the power spectral density for a cyclostationary process. Unfortunately, this is not true for a stationary process. For a white Gaussian noise process, the SCSD exhibits non-zero values for all cyclic frequencies although the noise process does not have any cyclic features. To ameliorate the challenge, the authors introduce noise elimination methods in the BCD. The simulation results show that the receiver operating characteristics of the BCD are considerably superior to the energy detection especially in signal-to-noise ratio fluctuation. With a moderately increased computational complexity, the BCD can be very beneficial for millimeter-wave cognitive radios with fast fading.

Feature Level Solution to Noise Robust Speech Recognition in the context of Tonal Languages

Performance of a speech recognition system is highly dependent on the operational environments. The mismatched ambient conditions have adverse impact on the performance of an Automatic Speech Recognition (ASR) system. The speech parameterization techniques for tonal speech recognition are different from those used for non-tonal speech recognition. It is due to the fact that tonal speech has two components – basic linguistic unit and tone. The basic linguistic unit with different tones convey different meanings. Therefore, the feature set used for tonal speech recognition must have the capability to representing both of them. Tone is determined by the fundamental frequency of the speech signal which is highly sensitive to noise. Since at the time of parameterization of the non-tonal speech recognition systems, these highly noise-sensitive tone related information are discarded, the traditional noise elimination methods used for non-tonal speech recognition fail to deliver robust performance in tonal speech recognition. In the present study, we have analyze the performance of different commonly used feature sets for noisy tonal speech recognition. Hidden Markov Model (HMM) based speech recognizer has been used for performance evaluation. Noise elimination techniques sub-band spectral subtraction and Wiener filter have been used for noise reduction and their relative performance have been evaluated.

Intelligent Machining Monitoring Using Sound Signal Processed With the Wavelet Method and a Self-Organizing Neural Network

A methodology was developed using the sound signal based on wavelet analysis and self-organizing neural network (NN) to monitor cutting accuracy in an extremely noisy machining process. Sawing experiments were conducted under different levels of feed speed, depth of cut, and rotation speed to monitor the longitudinal waviness of sawn samples using laser displacement sensors as an index of cutting accuracy and sawing deviation. The sound of the cutting and idling processes was recorded using a microphone. The acquired acoustic signals were pre-processed using the wavelet de-noising method for background noise elimination. As the signal still encompasses the low-frequency components corresponding to the idling process (machine motor, saw vibration, etc.), a systematic wavelet thresholding method was applied to the coefficients of the decomposed signal at different levels to discard the sound signal components associated with the idling process. Inverse wavelet transform was then applied to make a synthesized signal from the original one. Different features were extracted from the original and synthesized signals and used to train a self-organizing NN. Group method of data handling (GMDH) NN was utilized for predicting the waviness from the sensory features. The GMDH model trained with features extracted from the synthesized signal outperformed the one trained with the original signal features. The results suggested that employing the proposed wavelet-based methodology enables the acoustic signal to be used in monitoring the manufacturing processes in an extremely noisy environment. Self-organizing NN was shown to have a promising performance without facing the difficulties in finding the network optimal architecture, which is a typical challenge in the conventional backpropagation NNs.

Stiffness and Damping Identification for Asymmetric Building Frame With In-plane Flexible Floors

In most building structures, floors with sufficient in-plane stiffness exist and an assumption of rigid in-plane stiffness is valid. However, in some building structures, an assumption of rigid in-plane stiffness does not hold. A method of system identification (SI) for physical parameters (stiffness, damping) is proposed for three-dimensional (3D) building structures with in-plane flexible floors. The stiffness and damping parameters of each vertical structural frame in the 3D building structure are identified from the measured floor horizontal accelerations together with the stiffness and damping parameters of each floor. It is shown that a batch processing least-squares estimation method for many discrete time-domain measured data enables the direct identification of both the stiffness and damping parameters of each vertical structural frame and the stiffness and damping parameters of each floor. The proposed method possesses an advantage that all stiffness and damping parameters of vertical frames and horizontal frames (floors) can be identified simultaneously without search iteration. The accuracy and reliability of the proposed method are made clear by numerical simulations for measured data without noise and measured data with noise. A method of noise elimination is proposed to enhance the identification accuracy. Finally, experiments using a shaking table are conducted for the accuracy investigation of the proposed identification method. It is confirmed that the proposed identification method possesses a reliable ability to identify the stiffness and damping parameters for 3D building structures with in-plane flexible floors.

Bearing Fault Diagnosis Based on CEMDAN Energy Weighting Method

Bearing fault diagnosis is an important way to ensure the safe operation of equipment. However, the impact caused by some faults in the industrial field is often covered by environmental noise, the weak feature extraction method of vibration signal under strong noise interference is particularly important. In order to solve the problem that weak fault features can’t be extracted by single noise reduction or feature enhancement method under strong noise interference, an energy weighting method based on time-frequency spectrum analysis is proposed in this paper, which combines noise elimination and feature enhancement methods to extract weak impact features under strong noise background. Through the test verification of bearing inner ring fault, the method proposed in this paper has certain practical value in practical application.

Efficient Image De-Noising Technique Based on Modified Cuckoo Search Algorithm.

The image restoration has emerged as a very vital investigation technique in the domain of the image processing. The underlying motive behind the image restoration is devoted to the augmentation of the perceived visual impact of image so as to make it almost identical to the original image. A host of exploration approaches are now in vogues which are intended to steer clear of the noise, thereby regaining the images with original quality. In our earlier research, two distinct noise elimination methods like the (OGHP) and SURE shrinkage were effectively employed for the purpose of denoising, though the relative PSNR and SSIM efficiencies did not come up to the desired level. In the innovative approach envisaged in the document, at the outset, the noise is included by means of two processes like the salt and pepper and impulse noise. Subsequently, the pre-processing methods are performed with the able assistance of two novel filters such as the adaptive median filter and adaptive fuzzy switching. Thereafter, the preprocessed image is furnished to the succeeding function of noise elimination like the (OGHP) and SURE shrinkage. In the course of the OGHP noise elimination technique, the GHP constraints are optimized by employing the Cuckoo Search Algorithm. Thereafter, the noise-eliminated image is effectively estimated with the help of the Discrete Wavelet Transform (DWT). The consequential noiseless images are subjected to the image restoration procedure by efficiently employing the AGA approach. The cheering performance outcomes chant the success stories of the novel image restoration method, highlighting its superlative efficiency. Moreover, the efficacy of the innovative approach is assessed by means of a set of noise-polluted images and contrasted with the modern noiseless image restoration technique.

Interference reduction of high-energy noise for modal parameter identification of offshore wind turbines based on iterative signal extraction

An iterative noise extraction and elimination method is proposed, to solve the difficulty of modal parameter identification caused by contaminated high-energy components in measured signals. The approach is based on the idea that if only the high-energy noise is extracted and eliminated accurately, modal parameter identification can be improved significantly by using the remaining (filtered) signal. A theoretical development is that a general form of noises is considered, i.e., they can be purely harmonic or damped, which implies noises from machine vibration or other unknown sources can be taken into account. The advantage of this approach is that only one significant, relatively noisy component is extracted at each iteration, i.e., only the most reliable component is extracted each time, so more accurate high-energy noise elimination is expected. To demonstrate the accuracy of the proposed method, a numerical signal with one high-energy noisy component is used. Numerical results show that the approach can extract and eliminate high-energy noise at 2.1 Hz accurately by setting a small model order. A special case is the noisy component as a harmonic, which means the traditional Fourier transform can be applied. However, one can conclude that the proposed method outperforms the Fourier transform due to its limitation of fixed frequency resolution. To further investigate the effectiveness of the approach, an experiment on a steel offshore platform is conducted, and experimental results indicate that when measured raw data is used, only the first mode can be identified; if filtered data is used by employing the proposed method, the first two modes can be identified with improved damping ratio estimation. Finally, sea-test data from an offshore wind turbine in operating conditions are used, results show that the first mode with a frequency of 0.3592 Hz and a damping ratio of 0.0151 are successfully identified, while no reasonable parameters can be obtained by employing the stochastic subspace identification method, even when the used model order reaches 300.

Efficient fuzzy feature matching and optimal feature points for multiple objects tracking in fixed and active camera models

In computer vision, the multiple objects tracking play a vital challenging role. To solve the issues in this research field, various traditional techniques had been developed. In this paper, we consider the problem of tracking multiple persons in a dynamic environment (background) such as illumination changes and shadow moving. Notably, i) Estimating camera motion and ii) Multiple persons tracking are the two main phases involved in our proposed approach. In the first phase, the good features were extracted using both the SIFT features extraction steps and Gaussian noise elimination method. Instead of using the conventional SIFT-based matching method, we have introduced a new fuzzy matching method to create an adaptive matching zone (region). Using this, the two corresponding features from different frames can be matched perfectly. The brightness of a matching feature of interest indicates its size. Additionally, we use the magnitude and direction of the motion for accurate elimination of camera motion. In the second phase, the persons are tracked from the moving object by finding the optimal feature points and clustering of final points are made as the moving persons (objects). Experimental validation was performed on different challenging datasets and promising results are achieved by our proposed method compared to other existing methods.

How do ships pass through L-shaped turnings in the Singapore strait?

The value of ship Automatic Identification System (AIS) data coexists with data noises, whereas a tangible noise elimination method plays an important role in data mining. In this study, we first address the big ship AIS data de-noising issue for the 43 million records from 15 legs in the Singapore Strait, in which ship position, speed, and course data are checked and corrected by our proposed Method-I. Moreover, a winding number-based Method-II is developed to select data within irregularly restricted water areas. Based on these data, we particularly design a search-and-cut Method-III to filter out ships that pass through an “L” turning (two or more neighboring legs forming a curved channel) in this study.Speeds are normally distributed within a certain range according to the real AIS data in the Singapore Strait. Three types of waters (i.e., legs, “L” turnings and the whole strait) are compared. The results indicate that ship speeds in “L” turnings are more normally distributed than the speeds in legs and the entire strait. In the Singapore Strait, ship speeds slow down by 5.26%–14.4% to pass through “L” turnings approximately. Moreover, our tangible Method-IV of identifying ship decelerating processes from a large number of navigational data indicates that changes in ship speed of an “L” turning have the least relationship with ship types or lengths. However, we find that the longer the length of the ship, (1) the longer the decelerating time, (2) the smaller the deceleration, and (3) the smaller the course changing rate.

A snowfall noise elimination using moving object compositing method adaptable to natural boundary

In recent years, the use of surveillance cameras is increasingly recommended and they have been installed in many places. Snowy conditions at the time of an accident were associated with the problem that cars and accident circumstances become difficult to discern in images shot during snowfall. Previous techniques proposed methods for elimination of noise caused by snow using image shift or dedicated filters for the elimination of snowfall in video. However, these are associated with issues such as inability to cope with heavy snowfall or moving objects fading from view or being hard to discern. The present study proposes a method for snowfall noise elimination by extracting moving objects using the travel and the size of the moving object region between continuous frames, and compositing images while correcting for brightness in the background images. By distinguishing between falling snow and other moving objects, we can prevent objects other than snowfall becoming invisible. Using video of actual vehicles driving in snowy conditions for our experiments, we confirmed that snowfall noise can be eliminated without moving objects in the video becoming invisible. Furthermore, we confirmed that moving objects can be incorporated into the composited background images without any sense that they are out of place.

提高光谱匹配精度的散射噪声消除方法

提高光谱匹配精度的散射噪声消除方法

Power line interference noise elimination method based on independent component analysis in wavelet domain for magnetotelluric signal

With the urbanization in recent years, the power line interference noise in electromagnetic signal is increasing day by day, and has gradually become an unavoidable component of noises in magnetotelluric signal detection. Therefore, a kind of power line interference noise elimination method based on independent component analysis in wavelet domain for magnetotelluric signal is put forward in this paper. The method first uses wavelet decomposition to change single-channel signal into multi-channel signal, and then takes advantage of blind source separation principle of independent component analysis to eliminate power line interference noise. There is no need to choose the layer number of wavelet decomposition and the wavelet base of wavelet decomposition according to the observed signal. On the treatment effect, it is better than the previous power line interference removal method based on independent component analysis. Through the de-noising processing to actual magnetotelluric measuring data, it is shown that this method makes both the apparent resistivity curve near 50 Hz and the phase curve near 50 Hz become smoother and steadier than before processing, i.e., it effectively eliminates the power line interference noise.

Specialization improved nonlocal means to detect periodic impulse feature for generator bearing fault identification

It is significant to perform damage identification of wind turbine using running condition data for guaranteeing its safe operation. Because acquired condition data is usually mixed with heavy background noise, feature enhancement and noise elimination method is necessary for this task. Nonlocal means algorithm increase the signal to noise ratio without destroying the original frequency spectrum structure, which can reserve the useful information farthest and meanwhile eliminate noise. And it seems to be a possible powerful tool along with demodulation technique (under constant speed condition) or order tracking analysis method (under variable speed condition) for the damage identification. However, the actual application cases show that it would obtain some not entirely satisfying results when face strong background noise situation. So, the specialization improved nonlocal means method is developed for the damage identification of generator bearing. Based on the analyzing the essence characteristic of mechanical vibration signal, more reasonable ideas on the algorithm design such as neighborhood selection and variation of weighting function during nonlocal means denoising for this task are proposed. The effectiveness of specialization improved nonlocal means method is verified by fault identification cases study including variable speed condition.

Preoperative Three-Dimensional Diagnosis of Neurovascular Relationships at the Root Exit Zones During Microvascular Decompression for Hemifacial Spasm

Hemifacial spasm occurs when a blood vessel compresses against an area near the root exit zone of the facial nerve. Developments in diagnostic neuroimaging have allowed three-dimensional (3D) observation of artery and nerve locations, an effective aid for treatment selection. However, an accurate interpretation of the 3D data remains challenging because imaging representations of complex small vessels are drowned out by noise. We used a noise elimination method to analyze artery and nerve locations and to determine their 3D relationship. Fifteen patients treated for hemifacial spasm were included. Images fused from 3 modalities of magnetic resonance imaging, 3D computed tomography, and angiography were used as source images. Using the images, models of the nerve and candidate vessels were created and shown in 3D to observe how the arteries were compressing the nerve and to identify the portions of the offending vessels that were closest to the nerve. These preoperative results were then compared with operative field observations during surgery. 3D models of the unaffected side were created and evaluated as controls. We confirmed that these models were accurate reconstructions of the source images as the tubular nerve and artery cross-sections showed good alignment onto magnetic resonance imaging axial slice images. The preoperative diagnoses of the compression sites and offending arteries all matched intraoperative findings. An accurate identification of the offending arteries and compression sites was possible, and this method is anticipated to offer effective means of preoperative simulation.

Image-Based Motion-Tolerant Remote Respiratory Rate Evaluation

Contact measurements of the respiratory rate using conventional electrocardiogram equipment requires patients to wear chest straps that can cause skin irritation and discomfort. Therefore, a real-time robust non-contact technique is developed for the measurement of respiratory rate variation. The changes of a simple harmonic motion between inhalation and exhalation from the human’s upper body can be observed from visual appearance. Therefore, in this paper, to characterize the motion, a salient region is automatically selected in the energy map resulted from Haar-like features through consecutive frames. Furthermore, a median optical flow signal is used to acquire the primary respiratory rate signal. The effective respiratory frequencies resulted from vertical motion variation are decomposed by median motion signal and then characterized by zero-crossing method before the frequencies were rectified by an estimation using the noise elimination methods. In the experiment, the performance is evaluated using an extensive data set obtained under distinct four respiratory types, regular respiration, respiration with body motion, respiration with distinct poses, and respiration with distinct capturing distances. Our approach develops a convenient non-contact method to evaluate the respiratory rate with the achievement of measurement under farther distance and also outperforms the current state-of-the-art approach in terms of high correlation coefficient. Therefore, the experiment results show its efficacy for real-world environment.

Noise Elimination Method in Automatic Fire Detection Equipment in Accordance with the Communication Distance

This research focused on the communication noise level depending on the distance for conventional fire detection systems (RS-485 codetransmitter) in comparison to the newly developed optical cable automatic fire detection system (optical communication code-transmitter. The noise levels of the RS-485 code-transmitter were respectively 10, 200, 1200, and 2400 mV for the communication distances of 0, 1, 40, and 80 m and showed a correlation between the increase of the communication noise with the increase of the communication distance. However, the optical communication code-transmitter did not show any sign of communication noise at these four different distances. Therefore, the newly developed optical communication system is an ideal system to avoid false alarms in automatic fire detection system.

3D seismic denoising based on a low-redundancy curvelet transform

Contamination of seismic signal with noise is one of the main challenges during seismic data processing. Several methods exist for eliminating different types of noises, but optimal random noise attenuation remains difficult. Based on multi-scale, multi-directional locality of curvelet transform, the curvelet thresholding method is a relatively new method for random noise elimination. However, the high redundancy of a 3D curvelet transform makes its computational time and memory for massive data processing costly. To improve the efficiency of the curvelet thresholding denoising, a low-redundancy curvelet transform was introduced. The redundancy of the low-redundancy curvelet transform is approximately one-quarter of the original transform and the tightness of the original transform is also kept, thus the low-redundancy curvelet transform calls for less memory and computational resource compared with the original one. Numerical results on 3D synthetic and field data demonstrate that the low-redundancy curvelet denoising consumes one-quarter of the CPU time compared with the original curvelet transform using iterative thresholding denoising when comparable results are obtained. Thus, the low-redundancy curvelet transform is a good candidate for massive seismic denoising.