Filtering Algorithm Research Articles

Real-Time Multi-Person Video Synthesis with Controllable Prior-Guided Matting.

In order to enhance the matting performance in multi-person dynamic scenarios, we introduce a robust, real-time, high-resolution, and controllable human video matting method that achieves state of the art on all metrics. Unlike most existing methods that perform video matting frame by frame as independent images, we design a unified architecture using a controllable generation model to solve the problem of the lack of overall semantic information in multi-person video. Our method, called ControlMatting, uses an independent recurrent architecture to exploit temporal information in videos and achieves significant improvements in temporal coherence and detailed matting quality. ControlMatting adopts a mixed training strategy comprised of matting and a semantic segmentation dataset, which effectively improves the semantic understanding ability of the model. Furthermore, we propose a novel deep learning-based image filter algorithm that enforces our detailed augmentation ability on both matting and segmentation objectives. Our experiments have proved that prior information about the human body from the image itself can effectively combat the defect masking problem caused by complex dynamic scenarios with multiple people.

A novel least squares support vector machine-particle filter algorithm to estimate the state of energy of lithium-ion battery under a wide temperature range

The state of energy (SOE) is a key indicator for lithium-ion battery management systems (BMS). Based on the second-order resistance-capacitance equivalent circuit model and online parameter identification using the dynamic weights particle swarm optimization (DWPSO) method, a least-squares support vector machine-particle filter (LSSVM-PF) algorithm is proposed to construct a particle filter to estimate the SOE of a lithium-ion battery, and then transfer the resulting estimation error together with the experimentally measured voltage and current values to a trained LSSVM model, and use the LSSVM model to optimize the SOE estimates obtained by the PF algorithm twice to improve the accuracy of SOE estimation for lithium-ion batteries. The feasibility of the proposed algorithm is verified using two complex operating conditions and at three different temperatures. The validation results show that the maximum error of SOE estimation of the proposed algorithm is 0.0284 for a wide temperature range under Beijing Bus Dynamic Stress Test (BBDST) condition, and 0.0226 for a wide temperature range under Dynamic Stress Test (DST) condition. The proposed algorithm significantly improves the accuracy of SOE estimation and provides a reference for fundamental applications of lithium-ion batteries.

Multi-interest adaptive unscented Kalman filter based on improved matrix decomposition methods for lithium-ion battery state of charge estimation

Accurately predicting the state of charge (SOC) is crucial to improving Li-ion battery performance. However, available model-based estimation approaches still face challenges in handling model uncertainty and measurement noise effects on parameter identification. Besides, the widely used unscented Kalman filter (UKF) algorithm has limitations in electric vehicles as it requires the error covariance matrix to maintain positive definiteness, limiting its applicability under certain conditions. This study introduces the bias-compensated forgetting factor recursive least squares (BCFFRLS) method for parameter estimation within the second-order RC equivalent circuit model specific to the INR18650-20R battery. Furthermore, we propose a novel algorithm named the optimization multi-interest adaptive unscented Kalman filter (O-MIAUKF). This algorithm is designed to address stability and robustness issues with traditional UKF encounters in dynamic environments. Experimental validation demonstrates that the O-MIAUKF algorithm excels in maintaining strong stability and robustness in various working conditions, accurately estimating SOC even with a non-positive covariance matrix. The SOC estimation error remains stable at 0.8 %, which is lower than that of the current Extended Kalman Filter (EKF), UKF, and Dual Extended Kalman Filter (DEKF).

Sign Language Detection using ML Technologies

Sign language recognition is a leap forward for helping deaf-mute people. In sign language, every sign has a meaning assigned to it, so that it becomes easy to understand and interpret by the people. The main primary objective of our project is to bridge the gap between the deaf and dumb people and ordinary people without the need of an intermediary translator. Gesture recognition is a computer based visual technique used to detect and identify the object in an image or video. An application of this gesture recognition technique involves translating the sign language into the American language which can be further understood and interpreted by normal people. Many researchers have proposed their systems for the implementation of the ASL system. This report is a review of some studies related to the same topic. We addressed different approaches that used Convolutional Neural Networks (CNNs), K- Nearest Neighbours algorithm (KNN), Edge detection algorithm, Deep Neural Network (DNN), fuzzy clustering machine learning algorithm, Kernelized correlation filters (KCF) algorithm along with their results and drawbacks. Furthermore, we report on research gaps while summarizing these studies. From the above research papers, we got an average accuracy of 90.24%. To overcome the drawbacks, we propose a system which collects images for deep learning using webcam and OpenCV, with the help of TensorFlow Object Detection and Python that allows you to translate sign language in real time

FPGA Implementation of a Pipelined Kalman Filter for Object Tracking in Two Dimensions

In a location tracking system for a moving object, not only accuracy but also real-time processing are important factors. The Kalman filter, as a recursive function, stands out as one of the prominent algorithms for object tracking. It continuously compares measured data with predicted values based on the system characteristics in real time, and then corrects the error of the predicted values while considering the noise of both system and measured data. This paper focuses on designing a hardware-based Kalman filter for object tracking in two dimensions. Following an analysis of the Kalman filter algorithm, the blocks capable of parallel processing are identified and configured to be processed in parallel, effectively reducing data processing time. The clock speed is enhanced by using the pipeline technique. In addition, the time-sharing technique is applied to increase the utilization of hardware resources and reduce the area. Data was processed at 32-bit floating points to uphold accuracy comparable to software-implemented Kalman filters. The proposed Kalman filter architecture is designed using verilog HDL and then simulated in Synopsys VCS/Verdi. And the accuracy is verified by comparing results with a software-based Kalman filter designed using MATLAB. It was implemented using Zynq ZYNQ-7 ZC702 Programmable logic via Xilnix Vivado, and can operate at 33MHz. It takes a total of 44 clocks, or 1.32 us, to process one data. Therefore, it was confirmed that the designed Kalman filter hardware is suitable for real-time processing.

Online estimation of inlet contaminant concentration using Eulerian-Lagrangian method of fundamental solutions and Bayesian inference

The advection-diffusion equation is fundamental to modeling various transport phenomena, including the distribution of chemical species in surface or groundwater flow. In cases where the concentration at the source is unknown, inverse problem formulations are required to estimate the desired states by assimilating concentration monitoring data from specific points along the watercourse using a Bayesian approach. This paper proposes a combination of the Eulerian-Lagrangian method and the meshless method of fundamental solutions to solve the advection-diffusion equation. Moreover, the method was used as an evolution model for the sequential importance resampling particle filter algorithm to reconstruct the time-dependent inlet pollutant concentration in inverse problems. Numerical smooth and discontinuous inlet function results show that the particle filter - Eulerian-Lagrangian method of fundamental solutions combination can reconstruct inlet concentration time series.

Improvement and optimization of coal dust concentration detection technology: Based on the 3σ criterion and the kalman filtering composite algorithm

Due to the inherent time delay in dust concentration data transmission, dust reduction equipment is unable to respond according to the current moment's dust concentration. Hence, the focus of mine dust concentration detection technology extends beyond the current dust concentration to accurately predicting the next moment's dust concentration, allowing a response window for dust reduction equipment, leading to more optimal dust reduction effects. This necessitates the introduction of the Kalman filtering algorithm. Despite the traditional single Kalman filtering algorithm's ability to make certain dust concentration predictions, it is constrained by high-temperature, high-humidity underground environments and electromagnetic pulse signal interference from fan and coal mining machine operations, with an error rate reaching up to 40 %, seriously disrupting underground dust reduction equipment. However, the composite Kalman filtering algorithm incorporating the 3σ rule effectively eliminates bad values and removes instability from the underground environment and noise, rendering it applicable to most complex environments, providing accurate predicted concentrations for dust reduction equipment, and boosting overall universality. This study designed a fusion algorithm based on the Pauta criterion and Kalman filter to further optimize the dust concentration detection technology based on electrostatic induction. The sensor's model and measuring principle were first analyzed, and the initial calibration and precision testing of the sensor using the experimental system were completed. The dust concentration fluctuation characteristics were then modeled with Matlab, and the noise variance was estimated to process and filter the dust concentration data with the Pauta criterion, Kalman filter, and fusion algorithm respectively. The accuracy of the dust concentration sensor was tested again after applying the fusion algorithm, which proved to significantly enhance the stability and accuracy of the electrostatic induction-based dust concentration sensor. This study has critical implications for coal mine dust concentration monitoring and coal miner occupational health protection.

Real-time model correction using Kalman filter for Raman-controlled cell culture processes

Raman spectroscopy has found extensive use in monitoring and controlling cell culture processes. In this context, the prediction accuracy of Raman-based models is of paramount importance. However, models established with data from manually fed-batch cultures often exhibit poor performance in Raman-controlled cultures. Thus, there is a need for effective methods to rectify these models. The objective of this paper is to investigate the efficacy of Kalman filter (KF) algorithm in correcting Raman-based models during cell culture. Initially, partial least squares (PLS) models for different components were constructed using data from manually fed-batch cultures, and the predictive performance of these models was compared. Subsequently, various correction methods including the PLS-KF-KF method proposed in this study were employed to refine the PLS models. Finally, a case study involving the auto-control of glucose concentration demonstrated the application of optimal model correction method. The results indicated that the original PLS models exhibited differential performance between manually fed-batch cultures and Raman controlled cultures. For glucose, the RMSEP of manually fed-batch culture and Raman-controlled culture was 0.23 and 0.40 g·L-1. With the implementation of model correction methods, there was a significant improvement in model performance within Raman-controlled cultures. The RMSEP for glucose from updating-PLS, KF-PLS, and PLS-KF-KF was 0.38, 0.36, and 0.17 g·L-1, respectively. Notably, the proposed PLS-KF-KF model correction method was found to be more effective and stable, playing a vital role in the automated nutrient feeding of cell cultures.

Lane-line Detection based on Inverse Perspective Mapping and Kalman Filter

In real-time vision-based autonomous driving, lane-lines are difficult to identify due to shadows, road disrepair, reflection or other disturbances, as well as obstacles dynamic occlusion. In this paper, region of interest (ROI) is selected first. An algorithm based on inverse perspective mapping (IPM) is proposed to transform graphic images into flat images, after the operation of gray processing and image enhancement, binarization processing and secondary image enhancement are carried out for these images. Through the pixel statistics of image columns, The characteristic curves are extracted and lane-lines are fitted. At the same time, Kalman filter algorithm is proposed to predict the lane-line which are discontinuity, this improves the robustness and stability of the lane-line detection effectively. Finally, Experiments have been conducted on different scenes, the results show that the proposed algorithm own good detection efficiency.

Lidar signal processing method for atmospheric coherence length measurement based on the WD-ADMF

A denoising method applied to atmospheric coherent length lidar is proposed. Wavelet decomposition (WD) and the adaptive median filter (ADMF) are combined in this method. In this research, the effectiveness of the WD-ADMF has been verified through simulation and measurement. The results show that this filter algorithm, when applied to lidar data, improves the average peak signal-to-noise ratio (PSNR) and centroid error while maintaining data integrity such that the measurement of coherence length or the inference of C n 2 from coherence length more closely matches simulated truth and measured data.

Study on the estimation of the state of charge of lithium-ion battery

Lithium-ion battery (LIB) occupies a major position in rechargeable batteries. However, if the state of charge (SOC) estimation of LIB is not accurate, it will lead to the decline of battery life. SOC is often estimated by the extended Kalman filter (EKF) algorithm, but this method depends on the accuracy of model parameters. The sampling point interval time can affect the accuracy of SOC estimation. In this paper, the first-order RC model parameters are identified by two methods and the SOC is estimated by the EKF algorithm. Using this method, the battery parameters are received more accurate and the SOC estimation error is smaller. When the sampling interval is 0.03 s, the SOC estimation error is the smallest. And the second-order RC model can receive the same result. When the sampling interval is 0.03 s, the SOC estimation error of the 1RC-1 method is less than 1.4 %, the SOC estimation error of the 1RC-2 method is less than 0.017, and the root mean square error (RMSE) is 0.0087 and 0.01084. The conclusion is also used in the second-order RC model and contributes to the improvement of the SOC estimation accuracy. An effective method to estimate the battery SOC range is also proposed. This paper provides an effective method for obtaining more accurate SOC.

A vision imaging compensation algorithm for quadruped robot based on walking jitter

To solve the problem that the walking jitter of quadruped robots leads to the degradation of clarity of visual imaging, a quadruped robot visual imaging jitter compensation algorithm based on the theory of walking jitter is proposed. The D-H coordinate transformation method is used to establish the coordinate system of each joint of the leg. The kinetic equations of the leg are derived from the relationship between the rotational velocity and the moment of the leg joint, and the kinetic equilibrium equations of the quadruped robot body are established based on the spatial moment equilibrium theorem; the spring-mass model of the leg of the quadruped robot is used to construct the kinetic equations of the leg jittering, and the kinetic equations of the body jittering are derived using the moment equilibrium condition of the body center of gravity position and under the effect of the leg and body jitter to obtain the visual imaging device jitter quantity; finally, the tremor quantity is combined with the jitter quantity and rotation matrix to derive the walking jitter mathematical model of the quadruped robot visual imager, and the jitter compensation algorithm of quadruped robot visual imager is verified. The experimental results show that compared with the traditional Wiener filter algorithm for jitter compensation and the BP neural network jitter compensation algorithm, this algorithm improves the visual imaging by 10.8% and 3.3% in the two evaluation indexes of peak signal-to-noise ratio and structural similarity, respectively, and the de-jittering effect is better.

Kalman filter with impulse noised outliers: a robust sequential algorithm to filter data with a large number of outliers.

Impulse noised outliers are data points that differ significantly from other observations. They are generally removed from the data set through local regression or the Kalman filter algorithm. However, these methods, or their generalizations, are not well suited when the number of outliers is of the same order as the number of low-noise data (often called nominal measurement). In this article, we propose a new model for impulsed noise outliers. It is based on a hierarchical model and a simple linear Gaussian process as with the Kalman Filter. We present a fast forward-backward algorithm to filter and smooth sequential data and which also detects these outliers. We compare the robustness and efficiency of this algorithm with classical methods. Finally, we apply this method on a real data set from a Walk Over Weighing system admitting around 60 % of outliers. For this application, we further develop an (explicit) EM algorithm to calibrate some algorithm parameters.

The effect of hair removal and filtering on melanoma detection: a comparative deep learning study with AlexNet CNN.

Melanoma is the most aggressive and prevalent form of skin cancer globally, with a higher incidence in men and individuals with fair skin. Early detection of melanoma is essential for the successful treatment and prevention of metastasis. In this context, deep learning methods, distinguished by their ability to perform automated and detailed analysis, extracting melanoma-specific features, have emerged. These approaches excel in performing large-scale analysis, optimizing time, and providing accurate diagnoses, contributing to timely treatments compared to conventional diagnostic methods. The present study offers a methodology to assess the effectiveness of an AlexNet-based convolutional neural network (CNN) in identifying early-stage melanomas. The model is trained on a balanced dataset of 10,605 dermoscopic images, and on modified datasets where hair, a potential obstructive factor, was detected and removed allowing for an assessment of how hair removal affects the model's overall performance. To perform hair removal, we propose a morphological algorithm combined with different filtering techniques for comparison: Fourier, Wavelet, average blur, and low-pass filters. The model is evaluated through 10-fold cross-validation and the metrics of accuracy, recall, precision, and the F1 score. The results demonstrate that the proposed model performs the best for the dataset where we implemented both a Wavelet filter and hair removal algorithm. It has an accuracy of 91.30%, a recall of 87%, a precision of 95.19%, and an F1 score of 90.91%.

A multi-level power grid enhanced identity authentication data management platform based on filtering algorithms

In response to the optimal extraction of DCT coefficients in facial images, the author proposes a DCT coefficient extraction method based on discriminant analysis. Based on the discriminant analysis of DCT coefficients, the DCT coefficients with high discriminant values are selected as features. Comparing the DPA based discrete cosine coefficient selection method proposed by the author with the traditional Zigzag discrete cosine coefficient selection method, experiments were conducted on the ORL face database and the Yale face database, respectively. The recognition performance on the ORL face database was higher than that on the Yale face database, as the facial image expression and lighting changes in the ORL database were relatively few, making it suitable for extracting key features. In response to the problem that the speech parameter MFCC is greatly affected by noise and can only reflect the static characteristics of speech, the author extracted gamma pass filtering cepstrum coefficients with human auditory characteristics and gamma pass sliding differential cepstrum coefficients that can reflect the dynamic characteristics of speech based on gamma tone filters and sliding differential cepstrum. In the NUST603 speech database, under pure background, the recognition rate based on GFSDCC features reached 89.88%, and the recognition effect based on GFCC features was 87.52%, which is 4.66% and 2.36% higher than that based on MFCC features. In noisy environments, the average recognition rates of speaker recognition systems based on GFCC and GFSDCC are 56.06% and 59.07%, while the average recognition rates of speaker recognition systems based on MFCC speech features are 53.89%, 2.17% and 5.18% higher, respectively. The gain in this recognition effect comes from the characteristics of the auditory model, as the Gammatone filter effectively reflects the noise resistance of the human auditory system.

State-of-charge estimation of lithium-ion battery: Joint long short-term memory network and adaptive extended Kalman filter online estimation algorithm

With the widespread use of lithium-ion batteries, estimating the State of Charge (SOC) has been one of the most critical tasks in the battery management system (BMS). In this paper, a joint long short-term memory recurrent neural network (LSTM-RNN) and adaptive extended Kalman filter (AEKF) algorithm is proposed to achieve accurate SOC estimation. The proposed algorithm generates an initial SOC estimation value through a trained LSTM-RNN with a sliding time window signal sequence as input. The initial estimation value is then used as a feedback input to the AEKF. The results show that only a 30-s time series as the input of the LSTM-RNN can achieve satisfactory estimation results. Compared with traditional adaptive extended Kalman filter and unscented Kalman filter (UKF), the proposed algorithm has a better performance of estimation accuracy under various discharging conditions, especially when applied at low temperatures. Finally, the SOC estimation results under four operating conditions and various temperatures show that the maximum error is less than 2.5 %, in particular, the root mean square error (RMSE) and the mean absolute error (MAE) are below 1.01 % and 0.64 %. The proposed algorithm is an effective SOC estimation method with good robustness and high estimation accuracy.

Research on Combined Localization Algorithm Based on Active Screening-Kalman Filtering.

Real-time acquisition of location information for agricultural robotic systems is a prerequisite for achieving high-precision intelligent navigation. This paper proposes a data filtering and combined positioning method, and establishes an active screening model. The dynamic and static positioning drift points of the carrier are eliminated or replaced, reducing the complexity of the original Global Navigation Satellite System (GNSS) output data in the positioning system. Compared with the traditional Kalman filter combined positioning method, the proposed active filtering-Kalman filter algorithm can reduce the maximum distance deviation of the carrier along a straight line from 0.145 m to 0.055 m and along a curve from 0.184 m to 0.0640 m. This study focuses on agricultural robot positioning technology, which has an important influence on the development of smart agriculture.

Second-Order Central Difference Particle Filter Algorithm for State of Charge Estimation in Lithium-Ion Batteries

The estimation of the state of charge (SOC) in lithium-ion batteries is a crucial aspect of battery management systems, serving as a key indicator of the remaining available capacity. However, the inherent process and measurement noises created during battery operation pose significant challenges to the accuracy of SOC estimation. These noises can lead to inaccuracies and uncertainties in assessing the battery’s condition, potentially affecting its overall performance and lifespan. To address this problem, we propose a second-order central difference particle filter (SCDPF) method. This method leverages the latest observation data to enhance the accuracy and noise adaptability of SOC estimation. By employing an improved importance density function, we generate optimized particles that better represent the battery’s dynamic behavior. To validate the effectiveness of our proposed algorithm, we conducted comprehensive comparisons at both 25 °C and 0 °C under the new European driving cycle condition. The results demonstrate that the SCDPF algorithm exhibits a high accuracy and rapid convergence speed, with a maximum error which never exceeds 1.30%. Additionally, we compared the SOC estimations with both Gaussian and non-Gaussian noise to assess the robustness of our proposed algorithm. Overall, this study presents a novel approach to enhancing SOC estimation in lithium-ion batteries, addressing the challenges posed by the process itself and measurement noises.

A study of the attenuation stage of a global infectious disease.

Differences in control measures and response speeds between regions may be responsible for the differences in the number of infections of global infectious diseases. Therefore, this article aims to examine the decay stage of global infectious diseases. We demonstrate our method by considering the first wave of the COVID-19 epidemic in 2020. We introduce the concept of the attenuation rate into the varying coefficient SEIR model to measure the effect of different cities on epidemic control, and make inferences through the integrated adjusted Kalman filter algorithm. We applied the varying coefficient SEIR model to 136 cities in China where the total number of confirmed cases exceeded 20 after the implementation of control measures and analyzed the relationship between the estimated attenuation rate and local factors. Subsequent analysis and inference results show that the attenuation rate is significantly related to the local annual GDP and the longitude and latitude of a city or a region. We also apply the varying coefficient SEIR model to other regions outside China. We find that the fitting curve of the average daily number of new confirmed cases simulated by the variable coefficient SEIR model is consistent with the real data. The results show that the cities with better economic development are able to control the epidemic more effectively to a certain extent. On the other hand, geographical location also affected the effectiveness of regional epidemic control. In addition, through the results of attenuation rate analysis, we conclude that China and South Korea have achieved good results in controlling the epidemic in 2020.

Detection of Vehicle Number Plate and Speed Using Machine Learning

In an era marked by a significant upsurge in vehicular traffic and the resultant increase in road accidents, the need for advanced traffic management systems is paramount. This paper presents an innovative solution, "Detection of vehicle number plate and speed using machine learning" which aims to enhance road safety by addressing the persistent problem of over-speeding. The project leverages state-of-the-art technologies, including machine learning, deep learning, and computer vision, to develop a smart system that can detect speeding vehicles, recognize license plates, and record crucial data related to speed limit violations. The primary motivation for this endeavor is the escalating rate of road accidents, particularly in India, driven by speeding in areas susceptible to accidents. The system employs image segmentation, corner detection algorithms, filtering algorithms, and automatic vehicle plate recognition techniques to achieve its objectives. By using machine learning and deep learning models, the system accurately identifies license plates, segments characters on these plates, and recognizes them. In summary, this project represents a significant advancement in road safety and traffic management. By deploying an intelligent system that can proactively detect over-speeding vehicles and maintain comprehensive records, it strives to create a safer and more responsible driving environment, ultimately working towards the prevention of accidents and saving lives. Key Words: Automatic Number Plate Recognition, Vehicle Speed Detection, Machine Learning, Computer Vision, Traffic Management