scholarly journals Research on Target Tracking Algorithm Based on Siamese Neural Network

2021 ◽  
Vol 2021 ◽  
pp. 1-11
Author(s):  
Haibo Pang ◽  
Qi Xuan ◽  
Meiqin Xie ◽  
Chengming Liu ◽  
Zhanbo Li
Keyword(s):  
Neural Network ◽  
Target Tracking ◽  
Evaluation Criteria ◽  
Attention Mechanism ◽  
Tracking Algorithm ◽  
Temporal Attention ◽  
Tracking Accuracy ◽  
Current Frame ◽  
Tracking Process ◽  
Recognition Ability

Target tracking is a significant topic in the field of computer vision. In this paper, the target tracking algorithm based on deep Siamese network is studied. Aiming at the situation that the tracking process is not robust, such as drift or miss the target, the tracking accuracy and robustness of the algorithm are improved by improving the feature extraction part and online update part. This paper adds SE-block and temporal attention mechanism (TAM) to the framework of Siamese neural network. SE-block can refine and extract features; different channels are given different weights according to their importance which can improve the discrimination of the network and the recognition ability of the tracker. Temporal attention mechanism can update the target state by adjusting the weights of samples at current frame and historical frame to solve the model drift caused by the existence of similar background. We use cross-entropy loss to distinguish the targets in different sequences so that their distance in the feature domains is longer and the features are easier to identify. We train and test the network on three benchmarks and compare with several state-of-the-art tracking methods. The experimental results demonstrate that the algorithm proposed is superior to other methods in tracking effect diagram and evaluation criteria. The proposed algorithm can solve the occlusion problem effectively while ensuring the real-time performance in the process of tracking.

scholarly journals Target Adaptive Tracking Based on GOTURN Algorithm with Convolutional Neural Network and Data Fusion

2021 ◽  
Vol 2021 ◽  
pp. 1-12
Author(s):  
Zhengze Li ◽  
Jiancheng Xu
Keyword(s):  
Neural Network ◽  
Data Fusion ◽  
Convolutional Neural Network ◽  
Target Tracking ◽  
Network Structure ◽  
Attention Mechanism ◽  
Tracking Algorithm ◽  
Adaptive Tracking ◽  
Target Template ◽  
Overall Performance

With the advent of the artificial intelligence era, target adaptive tracking technology has been rapidly developed in the fields of human-computer interaction, intelligent monitoring, and autonomous driving. Aiming at the problem of low tracking accuracy and poor robustness of the current Generic Object Tracking Using Regression Network (GOTURN) tracking algorithm, this paper takes the most popular convolutional neural network in the current target-tracking field as the basic network structure and proposes an improved GOTURN target-tracking algorithm based on residual attention mechanism and fusion of spatiotemporal context information for data fusion. The algorithm transmits the target template, prediction area, and search area to the network at the same time to extract the general feature map and predicts the location of the tracking target in the current frame through the fully connected layer. At the same time, the residual attention mechanism network is added to the target template network structure to enhance the feature expression ability of the network and improve the overall performance of the algorithm. A large number of experiments conducted on the current mainstream target-tracking test data set show that the tracking algorithm we proposed has significantly improved the overall performance of the original tracking algorithm.

scholarly journals Multi-target tracking algorithm in intelligent transportation based on wireless sensor network

Open Physics ◽  
2018 ◽  
Vol 16 (1) ◽  
pp. 1000-1008 ◽  
Author(s):  
Yang Lei ◽  
Yuan Wu ◽  
Ahmad Jalal Khan Chowdhury
Keyword(s):  
Target Tracking ◽  
Information Exchange ◽  
Intelligent Transportation System ◽  
Tracking Error ◽  
Intelligent Transportation ◽  
Tracking Algorithm ◽  
Tracking Accuracy ◽  
Multiple Targets ◽  
Tracking Process ◽  
Multi Target Tracking

Abstract The traditional extended Kalman algorithm for multi-target tracking in the field of intelligent transportation does not consider the occlusion problem of the multi-target tracking process, and has the disadvantage of low multi-target tracking accuracy. A multi-target tracking algorithm using wireless sensors in an intelligent transportation system is proposed. Based on the dynamic clustering structure, the measurement results of each sensor are the superimposed results of sound signals and environmental noise from multiple targets. During the tracking process, each target corresponds to a particle filter. When the target spacing is relatively close to each other, each master node realizes distributed multi-target tracking through information exchange. At the same time, it is also necessary to consider the overlap between adjacent frames. Since the moving target speed is too fast, the target occlusion has the least influence on the tracking accuracy, and can accurately track multiple targets. The experimental results show that the proposed algorithm has a target tracking error of 0.5 m to 1 m, and the tracking result has high precision.

ROBUST SONIC LOG TRACKING USING A MULTI-RESOLUTION APPROACH

2021 ◽  
Author(s):  
Ting Lei ◽  
◽  
Michiko Hamada ◽  
Adam Donald ◽  
Takeshi Endo ◽  
...  
Keyword(s):  
Sudden Change ◽  
Processing Technique ◽  
Tracking Algorithm ◽  
Signal Loss ◽  
Tracking Accuracy ◽  
Signal Processing Technique ◽  
Waveform Data ◽  
Sonic Log ◽  
Tracking Process ◽  
Sonic Logs

Borehole acoustic logging is an acquisition method that is regarded as the most efficient and reliable method to measure subsurface rock elastic property. It plays an important role in both well construction and reservoir evaluation. The acquisition is carried out downhole by firing a transducer and then collecting waveforms at an array of receivers. A signal processing technique such as the slowness-time-coherence method is used to process array waveform data to resolve slownesses from different arrivals. To label these slowness values, a classification algorithm is then required to first determine if a primary (P) or a secondary (S) arrival exists or not, and then label out the existing ones at each depth of the entire logging interval to deliver continuous compressional and shear slowness logs. Such a process is referred as automatic sonic log tracking process. Clearly, it is of great importance to be able to track log as accurately as possible. Traditional approaches either use predefined slowness or arrival time boundary to distinguish them or treats slowness peaks in consecutive depths like “moving particles” and use a particle tracking algorithm to estimate their trace. However, such a tracking algorithm is often challenged by a sudden change in formation types at bed boundary, fine-scale heterogeneity, downhole logging noise, as well as unpredicted signal loss due to bad borehole shape or gas influx. Consequently, the tracking process is often a tricky task that requires heavy manual quality control and relabeling process, which poses significant bottleneck for a timely delivery of sonic logs for downstream petrophysical and geomechanical applications. In this paper, we propose a new physical based multi-resolution tracking algorithm that can improve the robustness of the tracking process. The new algorithm is inspired by the fact that different resolution sonic logs can sense different rock volumes and therefore response differently to a thin layer or an interval with bad borehole conditions. It works by grouping slowness-time peaks with different resolutions to form clusters, which are then tracked by the connecting with its neighboring depths. As different resolution slownesses are physically constrained by the convolution response of heterogeneous layers, the cluster-based multi-resolution tracking approach exhibits better logging depth continuity than the traditional single-resolution methods. Outliers due to noise can be confidently avoided. Finally, remaining gaps due to shoulder bed boundary can be patched by a convolution constrained optimization process from coherences from different resolutions. This new approach is therefore referred as a multi-resolution approach and can significantly improve sonic log tracking accuracy than the single resolution approach. This new algorithm has been tested on several sonic logging field data and demonstrates robust tracking performance of sonic P&S logs. Additionally, with the multi-resolution processing, sonic logs with different resolution can be reliably obtained and a high-quality high-resolution sonic log can also be automatically delivered, which can then be used to match resolution of other petrophysical logs for various types of interpretation.

Tracking algorithm of snowboard target in intelligent system

2020 ◽  
pp. 1-9
Author(s):  
Zhipeng Li ◽  
Xiaolan Li ◽  
Ming Shi ◽  
Wenli Song ◽  
Guowei Zhao ◽  
...  
Keyword(s):  
Target Tracking ◽  
High Speed ◽  
Intelligent System ◽  
Mean Shift ◽  
Target Position ◽  
Tracking Algorithm ◽  
Tracking Accuracy ◽  
Mean Shift Algorithm ◽  
Target Track ◽  
The Mean

Snowboarding is a kind of sport that takes snowboarding as a tool, swivels and glides rapidly on the specified slope line, and completes all kinds of difficult actions in the air. Because the sport is in the state of high-speed movement, it is difficult to direct guidance during the sport, which is not conducive to athletes to find problems and correct them, so it is necessary to track the target track of snowboarding. The target tracking algorithm is the main solution to this task, but there are many problems in the existing target tracking algorithm that have not been solved, especially the target tracking accuracy in complex scenes is insufficient. Therefore, based on the advantages of the mean shift algorithm and Kalman algorithm, this paper proposes a better tracking algorithm for snowboard moving targets. In the method designed in this paper, in order to solve the problem, a multi-algorithm fusion target tracking algorithm is proposed. Firstly, the SIFT feature algorithm is used for rough matching to determine the fuzzy position of the target. Then, the good performance of the mean shift algorithm is used to further match the target position and determine the exact position of the target. Finally, the Kalman filtering algorithm is used to further improve the target tracking algorithm to solve the template trajectory prediction under occlusion and achieve the target trajectory tracking algorithm design of snowboarding.

scholarly journals Research on the Particle Filter Single-Station Target Tracking Algorithm Based on Particle Number Optimization

2021 ◽  
Vol 2021 ◽  
pp. 1-8
Author(s):  
Lieping Zhang ◽  
Jinghua Nie ◽  
Shenglan Zhang ◽  
Yanlin Yu ◽  
Yong Liang ◽  
...  
Keyword(s):  
Particle Filter ◽  
Target Tracking ◽  
Real Time ◽  
Particle Number ◽  
Tracking Algorithm ◽  
Tracking Accuracy ◽  
Time Performance ◽  
Single Station ◽  
Tracking Model ◽  
Number Of Particles

Given that the tracking accuracy and real-time performance of the particle filter (PF) target tracking algorithm are greatly affected by the number of sampled particles, a PF target tracking algorithm based on particle number optimization under the single-station environment was proposed in this study. First, a single-station target tracking model was established, and the corresponding PF algorithm was designed. Next, a tracking simulation experiment was carried out on the PF target tracking algorithm under different numbers of particles with the root mean square error (RMSE) and filtering time as the evaluation indexes. On this basis, the optimal number of particles, which could meet the accuracy and real-time performance requirements, was determined and taken as the number of particles of the proposed algorithm. The MATLAB simulation results revealed that compared with the unscented Kalman filter (UKF), the single-station PF target tracking algorithm based on particle number optimization not only was of high tracking accuracy but also could meet the real-time performance requirement.

scholarly journals Two-Step Galileo E1 CBOC Tracking Algorithm: When Reliability and Robustness Are Keys!

2012 ◽  
Vol 2012 ◽  
pp. 1-14 ◽  
Author(s):  
Aleksandar Jovanovic ◽  
Cécile Mongrédien ◽  
Youssef Tawk ◽  
Cyril Botteron ◽  
Pierre-André Farine
Keyword(s):  
Tracking Algorithm ◽  
Tracking Accuracy ◽  
Multipath Mitigation ◽  
Binary Offset Carrier ◽  
Signal Tracking ◽  
Gps Navigation ◽  
Tracking Process ◽  
Galileo Satellite ◽  
Navigation Signals ◽  
Reliability And Robustness

The majority of 3G mobile phones have an integrated GPS chip enabling them to calculate a navigation solution. But to deliver continuous and accurate location information, the satellite tracking process has to be stable and reliable. This is still challenging, for example, in heavy multipath and non-line of sight (NLOS) environments. New families of Galileo and GPS navigation signals, such as Alternate Binary Offset Carrier (AltBOC), Composite Binary Offset Carrier (CBOC), and Time-Multiplex Binary Offset Carrier (TMBOC), will bring potential improvements in the pseudorange calculation, including more signal power, better multipath mitigation capabilities, and overall more robust navigation. However, GNSS signal tracking strategies have to be more advanced in order to profit from the enhanced properties of the new signals.In this paper, a tracking algorithm designed for Galileo E1 CBOC signal that consists of two steps, coarse and fine, with different tracking parameters in each step, is presented and analyzed with respect to tracking accuracy, sensitivity and robustness. The aim of this paper is therefore to provide a full theoretical analysis of the proposed two-step tracking algorithm for Galileo E1 CBOC signals, as well as to confirm the results through simulations as well as using real Galileo satellite data.

Efficient Moving Target Tracking Algorithm: Design and Implementation of the Sand-Table Tracking Algorithm

2013 ◽  
Vol 475-476 ◽  
pp. 1032-1039
Author(s):  
Jia Qi Li
Keyword(s):  
Target Tracking ◽  
Algorithm Design ◽  
Tracking Algorithm ◽  
Moving Target ◽  
Tracking Accuracy ◽  
Moving Target Tracking ◽  
Design And Implementation ◽  
Tracking Ability

Working on the design of a new algorithm :sand_table algorithm.The algorithm could work well in recognizing and tracking an single moving target shot by camera or in a video .The algorithm works simple with low operation cost.May used in tracking different object of many kinds.The algorithm imitate the the process of falling sands to Greatly enhance the tracking ability and tracking accuracy.

scholarly journals Target Tracking Algorithm Based on Adaptive Scale Detection Learning

Complexity ◽  
2021 ◽  
Vol 2021 ◽  
pp. 1-11
Author(s):  
Dawei Yang
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
Target Tracking ◽  
Sudden Change ◽  
Tracking Algorithm ◽  
Learning Network ◽  
Template Update ◽  
Deep Learning Network ◽  
Kernel Correlation ◽  
Correlation Filtering

In this paper, to better solve the problem of low tracking accuracy caused by the sudden change of target scale, we design and propose an adaptive scale mutation tracking algorithm using a deep learning network to detect the target first and then track it using the kernel correlation filtering method and verify the effectiveness of the model through experiments. The improvement point of this paper is to change the traditional kernel correlation filtering algorithm to detect and track at the same time and to combine deep learning with traditional kernel correlation filtering tracking to apply in the process of target tracking; the addition of deep learning network not only can learn more accurate feature representation but also can more effectively cope with the low resolution of video sequences, so that the algorithm in the case of scale mutation achieves more accurate target tracking in the case of scale mutation. To verify the effectiveness of this method in the case of scale mutation, four evaluation criteria, namely, average accuracy, cross-ratio accuracy, temporal robustness, and spatial robustness, are combined to demonstrate the effectiveness of the algorithm in the case of scale mutation. The experimental results verify that the joint detection strategy plays a good role in correcting the tracking drift caused by the subsequent abrupt change of the target scale and the effectiveness of the adaptive template update strategy. By adaptively changing the number of interval frames of neural network redetection to improve the tracking performance, the tracking speed is improved after the fusion of correlation filtering and neural network, and the combination of both is promoted for better application in target tracking tasks.

The effective model of transformation of ideological and political education in universities based on artificial intelligence

2021 ◽  
pp. 1-9
Author(s):  
Yu Zhang ◽  
Xuying Sun
Keyword(s):  
Neural Network ◽  
Artificial Intelligence ◽  
Bp Neural Network ◽  
Target Prediction ◽  
Tracking Algorithm ◽  
Ideological And Political Education ◽  
Political Education ◽  
Tracking Accuracy ◽  
Transmission Of Knowledge ◽  
Tracking Model

In the context of artificial intelligence, the path of knowledge transmission needs to be transformed. In essence, the transmission of knowledge and the transformation of information transmission methods are integrated. This paper studies the foreign object tracking algorithm, analyzes the error in the target tracking algorithm, and uses the BP neural network principle to modify the IMM algorithm. Aiming at the problem of low tracking accuracy when the target is maneuvering, this paper analyzes the linearization error of Kalman filter and builds a BP neural network to correct the tracking model of IMM. The model creates a target prediction training set and a test set, optimizes the parameters of the neural network, and conducts simulation experiments using MATLAB, which proved that the model had a higher accuracy in predicting the target trajectory of foreign objects. Therefore, the transformation of ideological and political teaching mode in colleges and universities can be realized, and the intelligent classroom of ideological and political education and intelligent communication have technical support.

Sign in / Sign up

Export Citation Format

Share Document