Denoising Filter Research Articles

Performance evaluation of 67 denoising filters in ultrasound images: A systematic comparison analysis

AbstractNoise corrupts ultrasound images and degrades spatial and contrast resolutions. Hence, it is challenging to characterize the lesions precisely using ultrasound images. The present study aims to evaluate 67 denoising filters and select the best one for ultrasound image denoising. Seven test images were synthesized to evaluate the performance of filters at three different noise levels. Eleven full‐reference quantitative image quality metrics (IQMs) were employed to evaluate the performance of the filters. A new filter evaluation method, Rank Analysis, was introduced and utilized at each noise level. The ten best filters with the smallest mean rank in all noise levels were defined for further analysis on real ultrasound images. The Rank Analysis was also employed for real ultrasound images, and filters were evaluated based on 14 IQMs (11 full‐reference and three no‐reference). Finally, the best filter was defined using the repeated measures analysis statistical test. According to the Rank Analysis results, the Spatial correlation (SCorr) filter obtained the best results with the mean rank scores±SD of 1 ± 0, which was significantly better than the other nine filters (p < 0.001). The second‐best results were achieved by three filters, Bitonic, most homogeneous neighborhood, and Lee diffusion (p < 0.05). We concluded that SCorr is the best filter for ultrasound image denoising. It can be used in the pre‐processing step before segmentation and diagnostic procedures. In addition, a new filter evaluation method, Rank Analysis, was introduced in this study, which is easy to use, fast, and provides reliable results. So, it can be used to evaluate newly developed filters in the future studies.

Design of Approximate Bilateral Filters for Image Denoising on FPGAs

This paper presents the hardware design of fast and low-cost denoising filters suitable to be exploited in the enabling technologies for Industry 5.0. A novel approximate computing strategy is introduced to reduce the computational complexity of the image denoising operation and to comply with real-time requirements. Firstly, it is demonstrated that the novel approximate approach can be helpfully exploited in the design of reconfigurable denoising filters able to reach image qualities as close as possible to the precise software counterparts. The reconfigurability leads to hardware architectures run-time adaptable to different levels of noise, whereas the adopted approximation strategy limits hardware resources and energy requirements. Quality tests, performed at various image and kernel sizes, and noise standard deviations, demonstrate that the approximate denoising approach presented here reaches PSNR and SSIM comparable with the precise denoise filtering. In comparison with state-of-the-art FPGA-based competitors, the novel filters reduce the resources requirements by up to 70%, achieve frame rates up to 35 times higher, and dissipate more than 45% lower power. When implemented within the XC7Z7020 FPGA device, a <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$5\times 5$ </tex-math></inline-formula> filter designed as proposed here denoises <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$512\times 512$ </tex-math></inline-formula> grayscale images using only 1689 LUTs, 2635 Flip-Flops and 32 DSPs. Moreover, it processes up to 926.8 frames per second, consumes just 63mW @ 244MHz and, with a noise standard deviation equal to 10, it achieves an average PSNR of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\sim 33$ </tex-math></inline-formula> dB with an average SSIM of ~0.86.

De-Noising Brain MRI Images by Mixing Concatenation and Residual Learning (MCR)

Brain magnetic resonance images (MRI) are used to diagnose the different diseases of the brain, such as swelling and tumor detection. The quality of the brain MR images is degraded by different noises, usually salt & pepper and Gaussian noises, which are added to the MR images during the acquisition process. In the presence of these noises, medical experts are facing problems in diagnosing diseases from noisy brain MR images. Therefore, we have proposed a de-noising method by mixing concatenation, and residual deep learning techniques called the MCR de-noising method. Our proposed MCR method is to eliminate salt & pepper and gaussian noises as much as possible from the brain MRI images. The MCR method has been trained and tested on the noise quantity levels 2% to 20% for both salt & pepper and gaussian noise. The experiments have been done on publically available brain MRI image datasets, which can easily be accessible in the experiments and result section. The Structure Similarity Index Measure (SSIM) and Peak Signal-to-Noise Ratio (PSNR) calculate the similarity score between the denoised images by the proposed MCR method and the original clean images. Also, the Mean Squared Error (MSE) measures the error or difference between generated denoised and the original images. The proposed MCR de-noising method has a 0.9763 SSIM score, 84.3182 PSNR, and 0.0004 MSE for salt & pepper noise; similarly, 0.7402 SSIM score, 72.7601 PSNR, and 0.0041 MSE for Gaussian noise at the highest level of 20% noise. In the end, we have compared the MCR method with the state-of-the-art de-noising filters such as median and wiener de-noising filters.

Novel Double Modular Redundancy Based Fault-Tolerant FIR Filter for Image Denoising

In signal processing and communication systems, digital filters are widely employed. In some circumstances, the reliability of those systems is crucial, necessitating the use of fault tolerant filter implementations. Many strategies have been presented throughout the years to achieve fault tolerance by utilising the structure and properties of the filters. As technology advances, more complicated systems with several filters become possible. Some of the filters in those complicated systems frequently function in parallel, for example, by applying the same filter to various input signals. Recently, a simple strategy for achieving fault tolerance that takes advantage of the availability of parallel filters was given. Many fault-tolerant ways that take advantage of the filter’s structure and properties have been proposed throughout the years. The primary idea is to use structured authentication scan chains to study the internal states of finite impulse response (FIR) components in order to detect and recover the exact state of faulty modules through the state of non-faulty modules. Finally, a simple solution of Double modular redundancy (DMR) based fault tolerance was developed that takes advantage of the availability of parallel filters for image denoising. This approach is expanded in this short to display how parallel filters can be protected using error correction codes (ECCs) in which each filter is comparable to a bit in a standard ECC. “Advanced error recovery for parallel systems,” the suggested technique, can find and eliminate hidden defects in FIR modules, and also restore the system from multiple failures impacting two FIR modules. From the implementation, Xilinx ISE 14.7 was found to have given significant error reduction capability in the fault calculations and reduction in the area which reduces the cost of implementation. Faults were introduced in all the outputs of the functional filters and found that the fault in every output is corrected.

A GPS Distance Error Forecast Model Based on IIR Filter Denoising and LSTM

Wireless channel prediction has attracted significant attention in recent years but is still considered an open topic. Global Positioning Systems (GPS) error prediction is one such example. Conventional mathematical and statistical models cannot predict time-varying channel variables within coherent time windows. Their mechanisms are geared towards minimizing mean square errors over all datasets rather than time-sequential prediction within coherent windows. On the other hand, Artificial Neural Network (ANN) based approaches have been introduced and studied for this purpose. ANNs have the capability to learn and identify underlying patterns in data. However, extracting and predicting wireless channel variables is a non-trivial task that presents challenges to researchers. Moreover, the widely applied Long Short-term Memory (LSTM) does not demonstrate satisfactory results because of unavoidable interferences, outliers and noise. To address these concerns, a novel LSTM with Infinite Impulse Response (IIR) filtering is proposed, which performs preprocessing of data smoothing. The proposed model reconfigures LSTM cells by adding new IIR gates, dedicated to removing noise and interferences for each cell in the hidden layers of ANN, and are self-adaptive through backpropagation during the training phase to achieve optimization. The model is evaluated with several commonly used time series forecast models, and the GPS error prediction results demonstrate that the proposed model is able to well adapt to the datasets with the best performance. The average prediction accuracy at the 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">th</sup> step is increased by 34% compared to the second-best model.

Comparative Study of Fingerprint-Based Gender Identification

Gender identification is a need in forensics investigation. In addition to their use for identification, fingerprints are used for gender identifications. In this work, we propose a model for fingerprint gender classification focusing on a small dataset with imbalanced classes. First, we applied denoising and equalization filters to the fingerprint images. Then, we cropped the image into a region of interest. After preprocessing the fingerprint image, we extracted fast Fourier transform (FFT) and principal component analysis (PCA) features, and we applied the min-max normalization. We train our models with each of them as well as with their fusion. To balance the initially imbalanced datasets, we used two sampling techniques: random sampling and synthetic minority oversampling Technique (SMOTE). Our comparison of the receiver operating characteristic (ROC) of K-nearest neighbours (KNN), Support Vector Machine (SVM), Decision Tree, and forward neural network showed that the best result for the area under the ROC Curve is given by the SVM classifier. Next, we compared the SVM classifier accuracy for each fingerprint position with FFT features, PCA features, and their fusion. In addition, we applied both random hybrid sampling and SMOTE sampling. We applied our method on the National Institute of Standards and Technology (NIST) special database, 4 gray scale images of fingerprint image. We showed that the fusion of FFT and PCA features gives the best gender identification accuracy when the SMOTE dataset balancing technique is applied. We found that the right ring finger was more appropriate with 75.55% of correct male gender identifications and 91.62% female gender identification. Further investigation into dataset balancing techniques will be applied in the future.

Long Distance Distributed Strain Sensing in OFDR by BM3D-SAPCA Image Denoising

We present a long distance distributed strain sensing in optical frequency domain reflectometry (OFDR) by shape-adaptive principal component analysis Block-Matching three-dimensional filter (BM3D-SAPCA) image denoising, which uses correlated patterns and high degree redundancy of sensing data for enhancing the performance of distributed sensing by image processing for removing noise and increasing the signal-to-noise ratio (SNR) of noisy measurements. Compared with 2D image denoising methods, BM3D-SAPCA method searches similar 2D image blocks and stacks them together in 3D arrays, which takes full advantage of the high level of similitude and redundancy contained on the multidimensional information to denoise. We find that the BM3D-SAPCA method can effectively suppresses noise aggravation along with an increasing of the sensing distance. Without modifying the hardware system of OFDR, we achieve a distributed strain sensing with a 5 cm spatial resolution, a 2 μϵ strain resolution on a 200 m all grating fiber. We compare the performance of distributed strain sensing by BM3D-SAPCA with Gaussian filter, wavelet denoising (WD) and non-local mean filter (NLM) using the same data. The mean maximal strain measurement error at loaded strain areas is reduced from 2.3791 μϵ to 0.6545 μϵ by BM3D-SAPCA. These mean errors by Gaussian filter, NLM and WD are 1.1177 μϵ, 1.6668 μϵ and 1.9721 μϵ, respectively. The mean standard deviations of strain measurement in eight repeat experiments is reduced from 1.5221 μϵ to 0.3134 μϵ after noise reduction by BM3D-SAPCA, which is 79.37% lower than the raw data. The mean standard deviations after noise reduction by Gaussian filter, NLM and WD are decreased by 68.85%, 64.24% and 14.41% respectively.

Performance analysis of various denoising filters on intravascular ultrasound coronary artery images

Accurate diagnosis of atherosclerotic coronary artery stenosis largely depends on intravascular ultrasound (IVUS) image quality. Coherent sources present during IVUS acquisition generates speckle noise and obstructs the clear view of the artery. Denoising aims to remove speckle noise while preserving edges and other important features. Accordingly, speckle noise suppression serves as an important preprocessing step for IVUS image analysis. This paper presents a performance analysis of various despeckling filters such as Median, Wiener, Speckle Reducing Anisotropic Diffusion (SRAD) filter, Wavelet, Non-local means (NLM) and curvelet based non-local means (CNLM) for IVUS images. To evaluate the efficacy, each filter is evaluated for three categories of cross-sectional IVUS images, that is, healthy, mild calcification, and dense calcification of real pullbacks from four patients. As it is challenging to denoise the IVUS images with complex lesions, that is, calcified arc >180°, in this performance analysis, the denoising performance for different denoising filters is considered not only for the healthy IVUS but also for complex lesions. Various measures such as peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), feature similarity index (FSIM), and expert evaluation are used to compare these approaches. The results depict that the CNLM denoising filter outperforms others in terms of PSNR and FSIM for speckle noise suppression for IVUS images while preserving the essential edge and feature details for better image visualization and diagnosis.

Analysis of Savitzky-Golay Filter for Electrocardiogram De-Noising Using Daubechies Wavelets

Electrocardiogram (ECG) examination is of great importance in medical diagnosis of the cardiac disease, but wrong interpretation due to noise interference in the signal could be dangerous as this may lead to wrong diagnoses of patient’s heart condition. De-noising helps to reduce the noise level for a better interpretation of the signals. In this study, an analysis of Savitzky-Golay (S-G) filter for ECG de-noising using Daubechies wavelets has been carried out using MATLAB version 2015a. Noisy ECG signals downloaded from physionet.org under MIT-BIH arrhythmia database was de-noised using S-G filter of polynomial order 9 to data frames of length 21 displayed in both time and frequency domains while a quantitative evaluation was carried out to check the performance of the filter under signal-to-noise ratio (SNR), mean square error (MSE) and signal-to-interference ratio (SIR). Results show that de-noising using S-G filter for SNR, MSE, and SIR gives an average value of 32.78dB, 0.0001 and 1852.358dB respectively. This implies that the S-G filter helps eliminates the background noise as well as maintaining a good fit for our data, and also do not allow co-channel interference from other radio transmitters, which makes it an excellent filter for ECG signal de-noising. Hospitals management and cardiac health centers most understand the importance of these parameters in the selection of de-noising filters for good quality ECG in diagnosis and treatment of cardiac patients.

Restoration of Lung Sound Signals Using a Hybrid Wavelet-Based Approach

A unique and ideal integration of wavelet-based total variation (WATV) and empirical Wiener denoising method is proposed in this article to significantly enhance the signal-to-noise ratio (SNR) while preserving the characteristics of a lung sound signal. While individual wavelet-based denoising filters based on a single basis function have been employed in the past, the outcome has been unsatisfactory because only significant (signal) wavelet coefficients are considered for denoising analysis. The new wavelet-based empirical Wiener (WATV-Wiener) hybrid technique, proposed here, takes into account both significant and insignificant (noise) wavelet coefficients of the noisy signal. An intensive analysis of selecting and fine-tuning the WATV-Wiener filter parameters is presented here through the simulation studies. The WATV-Wiener filter applied here onto different 1-D lung sound signals of different noise levels has led to an optimal root mean square error (RMSE) compared with seven other state-of-the-art filters reported in the literature. The optimal parameters achieved through our simulation studies led to a 3–20-dB improvement in SNR, and the average SNR was improved by 4–30 dB in our experiment. We also observed that the WATV-Wiener filter is less sensitive to the variation of SNR values of the input signal. Furthermore, the WATV-Wiener filter obtains similar SNR performance between continuous piecewise signal (wheeze) and noncontinuous piecewise signal (crackle) in both simulation and experimental studies.

Detection of a new crescent moon using the Maximally Stable Extremal Regions (MSER) technique

Visibility of a new crescent moon is crucial for Muslims when declaring the beginning of the Islamic calendar. For many years, astronomers had observed the visibility of the new crescent moon using various techniques, such as the naked eye and optical instruments. In line with technological developments, astronomers have changed their observation technique to using astronomical imaging with the aid of a digital camera attached to a telescope. This helps to capture and save the images of the new crescent moon. However, the new crescent moon image has presented difficulties to astronomers as it visibility in the image is difficult to detect due to a low contrast with the sky. Hence, the captured image could not be presented as evidence of its visibility. This study proposes the use of Contrast Limited Adaptive Histogram Equalization (CLAHE) for contrast enhancement. Then, the bilateral filter for image denoising was applied and lastly, the Maximally Stable Extremal Regions (MSER) detector was used to detect the new crescent moon region in the image. The result shows that this proposed method has acceptable practical results. The new crescent moon images can be enhanced successfully using CLAHE and the bilateral filter technique even when the new crescent moon is not visible to the naked eye. The new crescent moon region can be well detected with the MSER detector although MSER does not perform well on the blurry region of new crescent moon corner. This technique can help astronomers detect the new crescent moon’s visibility for declaring the beginning month of the Islamic calendar.

Low power design of energy efficient median filter

ABSTRACT The demand for cost, power and area-efficient de-noising filter is increasing day by day, because of its remarkable performance in various image processing applications. In this paper, a variable-node shift-based median filter is designed for achieving low power and area utilisation. Here, the proposed design encompasses a Max Tree Extraction (MTE) unit in conjunction with various devices, like comparators, variable-node shift generators, token generators, and noise detector units, which enable low hardware requirement and minimise the power consumption. The proposed model lessens the need of complex operations, like rank generation, rank calculation and median selection. Moreover, the proposed median filter is structured by Xilinx software and the hardware implementation is done by FPGA (Virtex 7 ×C7VX980 FPGA). Additionally, the proposed model is evaluated for different test images at various noise levels. The performance of the proposed median filter design is analysed under peak signal to noise ratio (PSNR), mean square error (MSE), power, delay and area. The experimental results show that the overall performance of the proposed method is superior to the state-of-art technologies.

Low-complexity trainable denoising filter structure for image recognition applications

Even in modern imaging equipment, different types of noise make their way into the images especially under bad lighting conditions and rough communication channels. While applying denoising filters, the preservation of the edge in images has always been an essential albeit a difficult task. This is particularly important for accuracy of further image processing tasks, such as recognition. We described a denoising approach to process noisy images specific for the task of classification with minimal impact on image resolution using low-complexity operations. Specifically, support vector machine-based machine learning classifiers have been proposed to dynamically select an appropriate filtering kernel while processing individual pixels in the input image. The selection is based on the image features in a local window. The provided experimental results on standard dataset prove that the proposed scheme performs as well as contemporary complex convolutional neural network (CNN)-based approaches for the task of denoising images before classification in the presence of a variety of mixed-type noise. Moreover, the proposed filtering approach can run at much higher speeds than their CNN-based counterparts while delivering similar or better performance. The proposed framework has been made publicly available as open-source code to facilitate further work by researchers and practitioners interested in this area.

Endorectal ultrasound radiomics in locally advanced rectal cancer patients: despeckling and radiotherapy response prediction using machine learning.

The current study aimed to evaluate the association of endorectal ultrasound (EUS) radiomics features at different denoising filters based on machine learning algorithms and to predict radiotherapy response in locally advanced rectal cancer (LARC) patients. The EUS images of forty-three LARC patients, as a predictive biomarker for predicting the treatment response of neoadjuvant chemoradiotherapy (NCRT), were investigated. For despeckling, the EUS images were preprocessed by traditional filters (bilateral, wiener, lee, frost, median, and wavelet filters). The rectal tumors were delineated by two readers separately, and radiomics features were extracted. The least absolute shrinkage and selection operator were used for feature selection. Classifiers including logistic regression (LR), K-nearest neighbor (KNN), support vector machine (SVM), random forest, naive Bayes, and decision tree were trained using stratified fivefold cross-validation for model development. The area under the curve (AUC) of the receiver operating characteristic curve followed by accuracy, precision, sensitivity, and specificity were obtained for model performance assessment. The wavelet filter had the best results with means of AUC: 0.83, accuracy: 77.41%, precision: 82.15%, and sensitivity: 79.41%. LR and SVM by having AUC: 0.71 and 0.76; accuracy: 70.0% and 71.5%; precision: 75.0% and 73.0%; sensitivity: 69.8% and 80.2%; and specificity: 70.0% and 60.9% had the highest model's performance, respectively. This study demonstrated that the EUS-based radiomics model could serve as pretreatment biomarkers in predicting pathologic features of rectal cancer. The wavelet filter and machine learning methods (LR and SVM) had good results on the EUS images of rectal cancer.

Active contour-based segmentation of normal and fetal spina bifida ultrasound images

Fetal spina bifida is a neurological disorder which occurs due to improper closure of the spinal column. Fetus identified with spina bifida suffers from various paralytic disorders throughout their lifespan. Early diagnosis of spina bifida aids in timely medical interventions. The ultrasound imaging is widely preferred for fetal monitoring. This study involves segmentation of the normal and abnormal fetal spine from ultrasound images using active contour algorithm. The images for analysis are collected from a diagnostic centre. The noise present in the images is removed using Wiener filter and anisotropic diffusion (AD) filter. The denoised images are evaluated with the metrics such as signal to noise ratio (SNR), peak signal to noise ratio (PSNR), structural similarity index measure (SSIM) and mean square error (MSE). The contrast enhancement is performed by histogram equalization (HE) and adaptive histogram equalization (AHE) techniques. The contrast enhanced images are validated by measures namely entropy and adaptive mean brightness error (AMBE). From the pre-processed image, the spine region is segmented using the active contour method. The results demonstrate that the AD filter with optimal parameters performs better than the Wiener filter for denoising. For the contrast enhancement, the AHE technique shows better performance compared to HE. The active contour technique is able to segment the spine regions in both the normal and spina bifida images. As early diagnosis of spina bifida is essential, this approach could be clinically significant.

Image enhancement techniques on deep learning approaches for automated diagnosis of COVID-19 features using CXR images.

The outbreak of novel coronavirus (COVID-19) disease has infected more than 135.6 million people globally. For its early diagnosis, researchers consider chest X-ray examinations as a standard screening technique in addition to RT-PCR test. Majority of research work till date focused only on application of deep learning approaches that is relevant but lacking in better pre-processing of CXR images. Towards this direction, this study aims to explore cumulative effects of image denoising and enhancement approaches on the performance of deep learning approaches. Regarding pre-processing, suitable methods for X-ray images, Histogram equalization, CLAHE and gamma correction have been tested individually and along with adaptive median filter, median filter, total variation filter and gaussian denoising filters. Proposed study compared eleven combinations in exploration of most coherent approach in greedy manner. For more robust analysis, we compared ten CNN architectures for performance evaluation with and without enhancement approaches. These models are InceptionV3, InceptionResNetV2, MobileNet, MobileNetV2, Vgg19, NASNetMobile, ResNet101, DenseNet121, DenseNet169, DenseNet201. These models are trained in 4-way (COVID-19 pneumonia vs Viral vs Bacterial pneumonia vs Normal) and 3-way classification scenario (COVID-19 vs Pneumonia vs Normal) on two benchmark datasets. The proposed methodology determines with TVF + Gamma, models achieve higher classification accuracy and sensitivity. In 4-way classification MobileNet with TVF + Gamma achieves top accuracy of 93.25% with 1.91% improvement in accuracy score, COVID-19 sensitivity of 98.72% and F1-score of 92.14%. In 3-way classification our DenseNet201 with TVF + Gamma gains accuracy of 91.10% with improvement of 1.47%, COVID-19 sensitivity of 100% and F1-score of 91.09%. Proposed study concludes that deep learning modes with gamma correction and TVF + Gamma has superior performance compared to state-of-the-art models. This not only minimizes overlapping between COVID-19 and virus pneumonia but advantageous in time required to converge best possible results.

A data-driven Gaussian process filter for electrocardiogram denoising

Gaussian Processes (GP)-based filters, which have been effectively used for various applications including electrocardiogram (ECG) filtering can be computationally demanding and the choice of their hyperparameters is typically ad hoc.We develop a data-driven GP filter to address both issues, using the notion of the ECG phase domain -- a time-warped representation of the ECG beats onto a fixed number of samples and aligned R-peaks, which is assumed to follow a Gaussian distribution. Under this assumption, the computation of the sample mean and covariance matrix is simplified, enabling an efficient implementation of the GP filter in a data-driven manner, with no ad hoc hyperparameters. The proposed filter is evaluated and compared with a state-of-the-art wavelet-based filter, on the PhysioNet QT Database. The performance is evaluated by measuring the signal-to-noise ratio (SNR) improvement of the filter at SNR levels ranging from -5 to 30dB, in 5dB steps, using additive noise. For a clinical evaluation, the error between the estimated QT-intervals of the original and filtered signals is measured and compared with the benchmark filter.It is shown that the proposed GP filter outperforms the benchmark filter for all the tested noise levels. It also outperforms the state-of-the-art filter in terms of QT-interval estimation error bias and variance.The proposed GP filter is a versatile technique for preprocessing the ECG in clinical and research applications, is applicable to ECG of arbitrary lengths and sampling frequencies, and provides confidence intervals for its performance.

Surface microseismic data denoising based on sparse autoencoder and Kalman filter

Microseismic technology is widely used in unconventional oil and gas production. Microseismic noise reduction is of great significance for the identification of microseismic events, the location of seismic sources and the improvement of unconventional oil and gas production. In this paper, a denoising filter is proposed based on sparse autoencoder and Kalman filtering. Firstly, a sparse autoencoder is pre-trained to learn the feature of the microseismic data. Sparse autoencoding is a back-propagation neural network algorithm based on unsupervised learning, in which there are three layers: the input layer, the hidden layer and the output layer. The hidden layer is the spare, which makes the algorithm learn features better, represents samples in harsh environments and reduces dimensionality effectively. Besides, Kalman filter is used to deal with the uncertainty factors. Using a dataset of 600 surface microseismic synthesis traces and simulation noise. Sparse autoencoders and Kalman filtering are trained to suppress noise. The denoising filter based on sparse autoencoder and Kalman filter model obtains a higher signal noise ratio than the conventional model. The experiment results for the filtering of surface microseismic signals show the feasibility and effectiveness of the proposed method.

Depth Image Denoising Algorithm Based on Fractional Calculus

Depth images are often accompanied by unavoidable and unpredictable noise. Depth image denoising algorithms mainly attempt to fill hole data and optimise edges. In this paper, we study in detail the problem of effectively filtering the data of depth images under noise interference. The classical filtering algorithm tends to blur edge and texture information, whereas the fractional integral operator can retain more edge and texture information. In this paper, the Grünwald–Letnikov-type fractional integral denoising operator is introduced into the depth image denoising process, and the convolution template of this operator is studied and improved upon to build a fractional integral denoising model and algorithm for depth image denoising. Depth images from the Redwood dataset were used to add noise, and the mask constructed by the fractional integral denoising operator was used to denoise the images by convolution. The experimental results show that the fractional integration order with the best denoising effect was −0.4 ≤ ν ≤ −0.3 and that the peak signal-to-noise ratio was improved by +3 to +6 dB. Under the same environment, median filter denoising had −15 to −30 dB distortion. The filtered depth image was converted to a point cloud image, from which the denoising effect was subjectively evaluated. Overall, the results prove that the fractional integral denoising operator can effectively handle noise in depth images while preserving their edge and texture information and thus has an excellent denoising effect.

The online detection system of concentric ring size based on machine vision

In today’s rapid industrial development, to improve production efficiency and product quality size detection process must be completed in real-time. Aiming at the problems of low production efficiency and large error rate caused by manual detection in the past, this paper proposes an online detection system of concentric ring size based on machine vision and develops a GUI software operation page. This system first processes the image gray level and uses the median filter for image denoising, and further uses Otsu threshold segmentation for image segmentation. Secondly, the Canny edge detection algorithm is used to extract the edge of the ring image. Finally, the radius and coordinates of the ring are measured by the improved Hough transform loop detection. Experiments show that the size measurement error of the measurement system is less than 0.015 mm. This system can meet the requirements of small and medium-sized enterprises for real-time detection of circular products, and improve the accuracy of quality inspection of circular products.