Background Estimation Algorithm Research Articles

Fault vibration model driven fault-aware domain generalization framework for bearing fault diagnosis

Deep learning methods can learn effective representations from the data, simplifying the fault diagnosis process and improving accuracy. However, the lack of data presents a significant challenge to effectively implementing fault diagnosis driven by deep learning. Simulation models can generate rich source domain data, which can be combined with domain adaptation methods to realize fault diagnosis tasks under sample scarcity. Nevertheless, there are some problems with existing simulation-driven fault diagnosis frameworks: (1) The choice of parameters for simulation models is usually empirical, which can lead to domain-specific components differences between the simulation data and the actual data. (2) It is difficult for the simulation models to generate the background noise of the real equipment, which will cause fault information to be covered, severely limiting the robustness of the fault diagnosis network. (3) Studies based on domain adaptation methods require a certain amount of target domain data for training, which is difficult to meet in real industrial production. To address these issues, this paper proposes a fault-aware domain generalization bearing fault diagnosis framework driven by fault vibration model, an analytical simulation model. Firstly, a spectral background estimation algorithm based on adaptive clutter separation (ACS) and maximum spectral envelope (MSE), termed ACS-MSE, is designed to estimate the potential resonance frequencies in the real system, guiding the parameters setting of the fault diagnosis vibration model and ensuring consistency between the simulated data and the actual data in the label space. Secondly, the normal actual signal is combined with the fault simulation signal to create a mixed signal that considers real background noise, and a fault-aware autoencoder is developed to extract the fault-related impact components covered by the background noise. Finally, a multi resonance frequencies domain generalization fault diagnosis network based on contrastive learning is designed to extract generalized fault features crossing different resonance frequencies and realize diagnosis tasks without real fault data. Experimental results on an open dataset and a custom dataset showed that the proposed method achieves higher accuracy compared to some advanced domain generalization and even domain adaptation methods. Ablation studies verify the importance and irreplaceability of the individual components of the proposed method.

Novel approaches for the estimation of the spectrum background for stationary and quasi-stationary signals

The vibration signals of rotating machinery contain information about the rotating components and the machine's structure. The peaks of a vibration spectrum are related to the vibration signals of the rotating components, and the background may be regarded as the spectrum without the “peaks”. In this paper, we extend studies on the estimation of the spectrum background. Two topics are addressed: first, an analysis of known methods and their extension into a new autonomic algorithm (autonomic Ceps-ACS) for background estimation of stationary signals, and second, a new approach for estimation of the background for quasi-stationary signals. The estimation of the background in both cases is based on two current techniques, namely, liftering of low quefrencies in the cepstrum domain (Ceps-Lift) and adaptive clutter separation (ACS). A relationship between the parameters of Ceps-Lift and ACS was formulated, enabling the development of the autonomic Ceps-ACS algorithm. For quasi-stationary signals, both Ceps-Lift and ACS have limited ability to estimate the spectrum background. To address the topic of estimation of the spectrum background of quasi-stationary signals, we proposed a novel approach that uses the background in the order domain to estimate the background in the frequency domain. Experimental measured transfer functions, measured data and simulated vibration signals were used to demonstrate the performance of the algorithms.

Comparative Monte Carlo Analysis of Background Estimation Algorithms for Unmanned Aerial Vehicle Detection

Background estimation algorithms are important in UAV (Unmanned Aerial Vehicle) vision tracking systems. Incorrect selection of an algorithm and its parameters leads to false detections that must be filtered by the tracking algorithm of objects, even if there is only one UAV within the visibility range. This paper shows that, with the use of genetic optimization, it is possible to select an algorithm and its parameters automatically. Background estimation algorithms (CNT (CouNT), GMG (Godbehere-Matsukawa-Goldberg), GSOC (Google Summer of Code 2017), MOG (Mixture of Gaussian), KNN (K–Nearest Neighbor–based Background/Foreground Segmentation Algorithm), MOG2 (Mixture of Gaussian version 2), and MEDIAN) and the reference algorithm of thresholding were tested. Monte Carlo studies were carried out showing the advantages of the MOG2 algorithm for UAV detection. An empirical sensitivity analysis was presented that rejected the MEDIAN algorithm.

A Defect Inspection for Explosive Cartridge Using an Improved Visual Attention and Image-Weighted Eigenvalue

Surface defects of explosive cartridge in the automatic sorting process are of a small area, irregular shape, and random distribution, and all problematic characteristics that hinder surface defect detection. To address these issues, a new multidefect detection method has been proposed in this paper based on a combination of an improved visual attention model and image partitioning-weighted eigenvalue (IP-WEV). First, image preprocessing is carried out by a background estimation algorithm. Then, a new fusion operator based on defects discrimination is implemented in a visual attention model to integrate intensity, orientation, and edge conspicuity into a saliency graph, in which a saliency effect of defects is considered during the fusion process. Third, a saliency map is divided into image blocks based on the image variance. This allows for the extraction of image blocks including defects, the calculation of the weighted eigenvalue, and the determination of regions containing multidefect. The IP-WEV is used to make a decision for multidefect. The experimental results show this method's detection accuracy as 98.2%, with a less computation time and quickly detection speed. Therefore, this method could be adopted for on-line detection systems for explosive cartridge surface defects.

Accelerating SuperBE with Hardware/Software Co-Design

Background Estimation is a common computer vision task, used for segmenting moving objects in video streams. This can be useful as a pre-processing step, isolating regions of interest for more complicated algorithms performing detection, recognition, and identification tasks, in order to reduce overall computation time. This is especially important in the context of embedded systems like smart cameras, which may need to process images with constrained computational resources. This work focuses on accelerating SuperBE, a superpixel-based background estimation algorithm that was designed for simplicity and reducing computational complexity while maintaining state-of-the-art levels of accuracy. We explore both software and hardware acceleration opportunities, converting the original algorithm into a greyscale, integer-only version, and using Hardware/Software Co-design to develop hardware acceleration components on FPGA fabric that assist a software processor. We achieved a 4.4× speed improvement with the software optimisations alone, and a 2× speed improvement with the hardware optimisations alone. When combined, these led to a 9× speed improvement on a Cyclone V System-on-Chip, delivering almost 38 fps on 320 × 240 resolution images.

SuperBE: computationally light background estimation with superpixels

This paper presents a motion-based superpixel-level background estimation algorithm that aims to be competitively accurate while requiring less computation time for background modelling and updating. Superpixels are chosen for their spatial and colour coherency and can be grouped together to better define the shapes of objects in an image. RGB mean and colour covariance matrices are used as the discriminative features for comparing superpixels to their background model samples. The background model initialisation and update procedures are inspired by existing approaches, with the key aim of minimising computational complexity and therefore processing time. Experiments carried out with a widely used dataset show that SuperBE can achieve a high level of accuracy and is competitive against other state-of-the-art background estimation algorithms. The main contribution of this paper is the computationally efficient use of superpixels in background estimation while maintaining high accuracy, reaching 135 fps on 320 × 240 resolution images.

An Improved Spectral Background Subtraction Method Based on Wavelet Energy.

Most spectral background subtraction methods rely on the difference in frequency response of background compared with characteristic peaks. It is difficult to extract accurately the background components from the spectrum when characteristic peaks and background have overlaps in frequency domain. An improved background estimation algorithm based on iterative wavelet transform (IWT) is presented. The wavelet entropy principle is used to select the best wavelet basis. A criterion based on wavelet energy theory to determine the optimal iteration times is proposed. The case of energy dispersive X-ray spectroscopy is discussed for illustration. A simulated spectrum with a prior known background and an experimental spectrum are tested. The processing results of the simulated spectrum is compared with non-IWT and it demonstrates the superiority of the IWT. It has great significance to improve the accuracy for spectral analysis.

Hybrid three-dimensional and support vector machine approach for automatic vehicle tracking and classification using a single camera

Using video in traffic monitoring is one of the most active research domains in the computer vision community. TrafficMonitor, a system that employs a hybrid approach for automatic vehicle tracking and classification on highways using a simple stationary calibrated camera, is presented. The proposed system consists of three modules: vehicle detection, vehicle tracking, and vehicle classification. Moving vehicles are detected by an enhanced Gaussian mixture model background estimation algorithm. The design includes a technique to resolve the occlusion problem by using a combination of two-dimensional proximity tracking algorithm and the Kanade–Lucas–Tomasi feature tracking algorithm. The last module classifies the shapes identified into five vehicle categories: motorcycle, car, van, bus, and truck by using three-dimensional templates and an algorithm based on histogram of oriented gradients and the support vector machine classifier. Several experiments have been performed using both real and simulated traffic in order to validate the system. The experiments were conducted on GRAM-RTM dataset and a proper real video dataset which is made publicly available as part of this work.

Motion detection based on the combining of the background subtraction and the structure–texture decomposition

To detect the moving objects in a video sequence based on background subtraction approaches, a background model should be estimated at the first time before subtract it from each image of the sequence and then segmenting the moving objects. In this paper, we present a new approach based on the combining of the background subtraction and the structure–texture decomposition (BS–STD). First, each gray-level image of the sequence will be decomposed on two components, structure and texture/noise by applying the Osher and Vese algorithm. The structure component of each image of the sequence will be taken to generate the background model using the median filter. The absolute difference used to subtracting the background before compute the binary image of the moving objects using the threshold generated by Otsu's method. The structure–texture decomposition (STD) is also used to ameliorate the results of methods, the background estimation algorithm with Σ−Δ (SDE), Simple Statistical Difference (SSD) and Motion detection with pyramid structure of background model (MDPS). The experimental results demonstrate that our approach is effective and accurate for moving objects detection and the structure–texture decomposition adds improvements to the results of state-of-the-art methods.

Star extraction based on clustering within a star tracker

The traditional star extraction methods of star tracker need much processing time and large image memory, resulting in slow attitude acquisition and complex electronic circuits. A new star extraction method, star extraction based on clustering, was developed. The analysis of on-line background estimation algorithm and clustering algorithm in the new method was introduced. The method was realized in one of the novel star trackers developed by Beijing Institute of Control Engineering (BICE), on-board circumlunar return and reentry spacecraft of 3rd phase of China lunar exploration program. According to the house-keeping data packages recorded during on-ground real sky test and the flight mission, the processing time of star extraction was reduced by 84% at least and the image memory was reduced by 99% at least, compared with traditional methods. The results indicated that the star extraction based on clustering had advantages in processing time, memory and migration,which made it possible to realize the method in almost all kinds of optical sensors.

Smoke Detection Using GMM and Adaboost

Detection performance of the smoke in the outdoor depends on weather conditions, the shadow by the movement of the sun, or illumination changes, etc. In this paper, a smoke detection system in conjunction with a robust background estimate algorithm to environment change in the outdoor in daytime is proposed. Gaussian Mixture Model (GMM) is applied as background estimation algorithm, and also, Adaboost algorithm is applied to detect the smoke for candidate region. Through the experiments with input videos obtained from a various weather conditions at the same actual road, the proposed algorithms were useful to detect smoke in the outdoor. Index Terms—GMM, adaboost, smoke detection, outdoor environments.

실외 도로에서의 영상기반 차량 감시에 관한 연구

실외 도로에서의 차량 검출 성능은 기상 상태, 태양 이동에 의한 그림자, 조도 변화 등에 영향을 받는다. 본 논문에서는 낮 시간대의 실외도로에서 이러한 주변 환경변화에 강건한 배경 추정 알고리즘과 연동한 차량 검출 시스템을 제안한다. 배경 추정 알고리즘은 혼합 가우시안 모델을 적용하고 후보 영역에 대한 차량 검출은 Adaboost 알고리즘을 적용하였다. 흐린 날, 비오는 날 등 동일한 실제 도로에서 서로 다른 기후에 획득한 CCTV 비디오 영상을 사용한 실험을 통해 제안하는 방법이 일반 도로에서의 차량 검출에 유용한 것을 확인하였다. Detection performance of the vehicle on the road depends on weather conditions, the shadow by the movement of the sun, or illumination changes, etc. In this paper, a vehicle detection system in conjunction with a robust background estimate algorithm to environment change on the road in daytime is proposed. Gaussian Mixture Model is applied as background estimation algorithm, and also, Adaboost algorithm is applied to detect the vehicle for candidate region. Through the experiments with input videos obtained from a various weather conditions at the same actual road, the proposed algorithm were useful to detect vehicles in the road.

Kernel-Based Machine Learning for Background Estimation of NaI Low-Count Gamma-Ray Spectra

Virtually all gamma-ray spectrometry measurements contain background components due to the ubiquitous presence of primordial radionuclides in the Earth's crust and cosmic radiation interactions high in the Earth's atmosphere. In principle, spectral signatures due to radiation source(s) of actual interest can be extracted from the measured gamma-ray spectrum by background subtraction. However, if separate background measurements are unavailable or infeasible, and particularly for measurements exhibiting low signal-to-noise ratio (SNR), background subtraction is nontrivial, and it requires accurate background estimation . An example application of gamma-ray spectroscopy with low SNR is the “source search” scenario, where the position of a source is sought using measurements taken over very short time intervals by a detector in motion. We have developed an algorithm for background estimation in low-count gamma-ray spectra using kernel-based Gaussian processes (GP) taken from the field of machine learning. We have evaluated the performance of our algorithm using a group of three kernels tested against a dataset composed of background spectra measured in an urban environment using a mobile sodium iodide (NaI) detector. We have also simulated datasets containing nonbackground gamma-ray sources in an urban background measured with a NaI detector. The simulated scenarios employed a variety of source-detector distances and different types of source shielding. As a metric of algorithm performance, we calculated correlation coefficients, Theil inequality coefficients, and count difference statistics between estimated and actual backgrounds. We concluded that our method adequately estimates the gamma-ray background, but we also observed a strong dependence of the algorithm's performance on the selected kernel.

Optimal criterion to select the background estimation algorithm for detection of dim point targets in infrared surveillance systems

A well-established scheme for target detection in infrared (IR) surveillance systems consists of applying a suitable decision rule on the images with background clutter previously removed. Background removal is accomplished by subtracting, from the original image, the estimate of the spatially varying background signal obtained by a background estimation algorithm (BEA). The overall target detection performance is strongly influenced by the effectiveness of the employed BEA. Particularly, the BEA and its design parameters should be chosen so as to get an accurate estimate of the background signal and to avoid biases caused by the possible presence of targets (target leakage). In this work, we present a novel method for the choice and setting of the best performing background removal technique for the detection of dim point targets. The proposed procedure is based on the simulation of dim targets implanted on an acquired sample image representing the scenario of interest. The choice of the best performing BEA is made by exploring the performance of the detection scheme for several configurations of the characteristic parameters of the BEAs. The effectiveness of the BEA selection procedure is evaluated in two case studies where real image sequences acquired by IR cameras are employed. The results confirm the benefits introduced by the proposed technique. Indeed, the performance of the IR detection system with the BEA tuned according to the proposed selection criterion is improved in that the number of false alarms is reduced up to 2 orders of magnitude compared with BEAs in other common configurations.

Estimation of In Situ 3-D Particle Distributions From a Stereo Laser Imaging Profiler

In this paper, an image processing system for estimating 3-D particle distributions from stereo light scatter images is described. The system incorporates measured, three-component velocity data to mitigate particle blur associated with instrument motion. An iterative background estimation algorithm yields a local threshold operator that dramatically reduces bias in particle counts over the full image field. Algorithms are tested on simulated particle distributions and data from an open-ocean profile collected near the Santa Barbara Channel Islands, CA. They yield over a 50% reduction in root-mean-squared error in particle size estimates, and a 30% reduction in the magnitude of the motion blur point spread function. In situ particle distributions are estimated and compared to several models. It is demonstrated that quantitative, 3-D particle distributions can be accurately estimated from these data for particles with diameter larger than 4 pixels (0.8 mm).

A NEW ADAPTIVE LINEAR COMBINED CFAR DETECTOR IN PRESENCE OF INTERFERING TARGETS

In this paper, a new radar constant false alarm rate detector to perform adaptive threshold target detection in presence of interfering targets is proposed. The proposed CFAR detector, referred to as Adaptive Linear Combined CFAR, ALC-CFAR, employs an adaptive composite approach based on the well-known cell averaging CFAR, CA-CFAR, and the ordered statistics, OS-CFAR, detectors. Data in the reference window is used to compute an adaptive weighting factor employed in the fusion scheme. Based on this factor, the ALC- CFAR tailors the background estimation algorithm. The conducted Monte Carlo simulation results demonstrate that the proposed detector provides low loss CFAR performance in an homogeneous environment and also performs robustly in presence of interfering targets. The performances of the ALC-CFAR detector have been evaluated and compared with that of the CA-CFAR and the OS-CFAR detectors. The obtained results are presented and discussed in this paper.

A Low-Complexity Algorithm for Static Background Estimation from Cluttered Image Sequences in Surveillance Contexts

For the purposes of foreground estimation, the true background model is unavailable in many practical circumstances and needs to be estimated from cluttered image sequences. We propose a sequential technique for static background estimation in such conditions, with low computational and memory requirements. Image sequences are analysed on a block-by-block basis. For each block location a representative set is maintained which contains distinct blocks obtained along its temporal line. The background estimation is carried out in a Markov Random Field framework, where the optimal labelling solution is computed using iterated conditional modes. The clique potentials are computed based on the combined frequency response of the candidate block and its neighbourhood. It is assumed that the most appropriate block results in the smoothest response, indirectly enforcing the spatial continuity of structures within a scene. Experiments on real-life surveillance videos demonstrate that the proposed method obtains considerably better background estimates (both qualitatively and quantitatively) than median filtering and the recently proposed "intervals of stable intensity" method. Further experiments on the Wallflower dataset suggest that the combination of the proposed method with a foreground segmentation algorithm results in improved foreground segmentation.

Linear model for fast background subtraction in oligonucleotide microarrays

BackgroundOne important preprocessing step in the analysis of microarray data is background subtraction. In high-density oligonucleotide arrays this is recognized as a crucial step for the global performance of the data analysis from raw intensities to expression values.ResultsWe propose here an algorithm for background estimation based on a model in which the cost function is quadratic in a set of fitting parameters such that minimization can be performed through linear algebra. The model incorporates two effects: 1) Correlated intensities between neighboring features in the chip and 2) sequence-dependent affinities for non-specific hybridization fitted by an extended nearest-neighbor model.ConclusionThe algorithm has been tested on 360 GeneChips from publicly available data of recent expression experiments. The algorithm is fast and accurate. Strong correlations between the fitted values for different experiments as well as between the free-energy parameters and their counterparts in aqueous solution indicate that the model captures a significant part of the underlying physical chemistry.

Practical background estimation for mosaic blending with patch-based Markov random fields

In this paper, we present a new background estimation algorithm which effectively represents both background and foreground. The problem is formulated with a labeling problem over a patch-based Markov random field (MRF) and solved with a graph-cuts algorithm. Our method is applied to the problem of mosaic blending considering the moving objects and exposure variations of rotating and zooming camera. Also, to reduce seams in the estimated boundaries, we propose a simple exposure correction algorithm using intensities near the estimated boundaries.

MO‐E‐L100J‐06: Slot Scan Imaging with a High Frame Rate Flat Panel Detector‐Measurement and Correction for In‐Slot Scatter

Purpose: To report on the use of a new scatter measurement technique to characterize the scatter rejection properties and to further correct for the in‐slot scatter in slot scan imaging with a high frame rate flat panel detector. Method and Materials: A new slot scan imaging technique was developed with a high frame rate (7.5 fps) flat panel detector. 300 images were acquired as an x‐ray fan beam, defined by a slot collimator placed in front of the tube, was scanned across an anthropomorphic chest phantom. Time‐intensity curves were formed from the image sequence for all pixels. They were integrated over the fan beam passing frames to form the scatter rejected image signals and over all frames to form the open field image signals. By assuming the primary component to be constant during the fan beam exposure, the scatter components were extracted from the time‐intensity curves to estimate the in‐slot scatter using an iterative background estimation algorithm. The results were used to measure the scatter signals and to correct for in‐slot scatter. Results: Slot scan images of an anthropomorphic chest phantom were acquired and formed for various projected fan beam widths. 2D maps of scatter‐to‐primary ratios and contrast‐to‐noise ratio (CNR) improvement factors were generated and plotted for the slot‐scan images and compared with those for the full field, non‐grided images . In the lungs, CNR was improved by 30 to 50%, and in the mediastinum, CNR was improved by 60 to 150%. Conclusion: We have described and demonstrated the use of a new scatter measurement technique to estimate and correct for the in‐slot scatter on a pixel‐by‐pixel basis in slot scan imaging with a high frame rate flat panel detector.This work was supported in part by research grants EB00117 from the NIH‐NIBIB and CA51248 and CA104759 from the NIH‐NCI.