Low Localization Accuracy Research Articles

Improved Field-Based Soybean Seed Counting and Localization with Feature Level Considered.

Developing automated soybean seed counting tools will help automate yield prediction before harvesting and improving selection efficiency in breeding programs. An integrated approach for counting and localization is ideal for subsequent analysis. The traditional method of object counting is labor-intensive and error-prone and has low localization accuracy. To quantify soybean seed directly rather than sequentially, we propose a P2PNet-Soy method. Several strategies were considered to adjust the architecture and subsequent postprocessing to maximize model performance in seed counting and localization. First, unsupervised clustering was applied to merge closely located overcounts. Second, low-level features were included with high-level features to provide more information. Third, atrous convolution with different kernel sizes was applied to low- and high-level features to extract scale-invariant features to factor in soybean size variation. Fourth, channel and spatial attention effectively separated the foreground and background for easier soybean seed counting and localization. At last, the input image was added to these extracted features to improve model performance. Using 24 soybean accessions as experimental materials, we trained the model on field images of individual soybean plants obtained from one side and tested them on images obtained from the opposite side, with all the above strategies. The superiority of the proposed P2PNet-Soy in soybean seed counting and localization over the original P2PNet was confirmed by a reduction in the value of the mean absolute error, from 105.55 to 12.94. Furthermore, the trained model worked effectively on images obtained directly from the field without background interference.

Denoising Method of Pipeline Leakage Signal Based on VMD and Hilbert Transform

In order to solve the problem of low localization accuracy of water supply pipeline leakage detection using acoustic emission signals, a noise reduction algorithm based on variational mode decomposition (VMD) and Hilbert transform is proposed in this paper. Firstly, the leakage signal is decomposed into several intrinsic mode functions (IMF) using VMD, and the number of decomposition layers is determined, and the IMF component is transformed by Hilbert to get the marginal spectrum, and the noise are preliminarily filtered according to the marginal spectrum characteristics and the cross‐correlation coefficient. Secondly, the signal denoising is realized by using filtering technology and cross‐correlation coefficient. Finally, the time delay estimation of the denoised signal is estimated and the location of the leakage point is calculated based on the positioning principle. In order to verify the performance of the algorithm, a water supply pipeline experiment platform for leakage simulation and leak location experiment was conducted. The cross‐correlation algorithm, VMD‐correlation coefficient algorithm, and the proposed algorithm are analyzed through simulation and experiment. The results show that compared with the other two algorithms, the proposed algorithm can effectively remove the noise in the leakage signal and improve the accuracy of leakage location.

VSC-HVDC Fault Location Method Based on Effective Feature and Depth Belief Network

Aiming at the problems that most fault location methods are applied to the high voltage direct current (HVDC) transmission system, such as low location accuracy and vulnerability to noise, a voltage source converter (VSC)-HVDC fault location method based on effective feature and deep belief network is proposed. Firstly, the Karenbauer phase mode transform is used to decouple the original fault current data to eliminate the coupling effect between positive and negative lines. Then, Hilbert Huang (HH) transform is used to obtain six effective features of information from frequency domain and time domain analysis to fully characterize the system fault current signal. Finally, the extracted effective features are input into the depth belief network for learning, so as to estimate the fault location in the transmission line. VSC-HVDC system is built based on the MATLAB platform for experimental analysis. The results show that the fault location obtained by the proposed method has the smallest overall fault location error, and the maximum and minimum errors are divided into about 0.52 km and -0.95 km. Compared with other methods, it has certain advantages in fault location.

Improved Particle Swarm Localization Algorithm Based on Evolutionary Mechanism

Aiming at the problem of low localization accuracy and poor robustness when conventional localization algorithms are used for the localization of WSN nodes, an improved particle swarm localization algorithm based on an evolutionary mechanism is proposed. To reduce the ranging error in a complex environment, the anchor box algorithm is improved by using ranging information, and the search area is reduced by the improved algorithm to determine the possible region where unknown nodes exist. Then the objective function is established to transform the localization problem into a bounded optimization problem. The conventional particle swarm optimization algorithm is improved and genetic operators such as crossover and mutation are introduced to improve the search accuracy and global optimization ability of the algorithm. The simulation results show that, compared with the conventional positioning algorithm, the proposed algorithm has smaller positioning errors and stronger robustness.

Simultaneous Localization of Scatterers and Target User Based on Indoor Prior Information in NLOS Environments

In complex indoor environments, there is often no Line of Sight (LOS) path between the transmitter and the receiver due to walls, moving people, and many other obstacles. Traditional localization method based on LOS measurement is no longer available which provides lower localization accuracy in Non-Line of Sight (NLOS) environments. In response to this problem, this paper proposes a joint localization algorithm for scatterers and target user based on indoor prior information in NLOS scenarios. The algorithm makes use of the prior localization information of the access point (AP) and the wall to determine the position of the scatterers and target user simultaneously. Since the established constrained optimization problem is non-convex which is a difference of convex (DC) problem, it is transformed into a convex one based on a Taylor series expansion (TSE) method before it is solved. At the same time, as only a single AP signal can be received normally, the proposed method is then extended to the scenario with only one AP. Simulation and real-world testing results both show that not only the proposed method based on multiple APs, but also the method based on a single AP achieve high positioning accuracy, which are better than the existing positioning methods mentioned in this article as well.

Research on Mobile Node Localization Optimization Algorithm in Wireless Sensor Networks Based on Monte Carlo Method

Mobile node localization is one of the key technologies in wireless sensor network applications. Aiming at the shortcomings of Monte Carlo algorithm (MCL) in mobile wireless sensor network node localization in nonideal environments, such as low accuracy and low sampling rate, in the prediction phase and filtering phase of MCL, the communication radius of the unknown node is determined according to the size of the irregularity of the node. Perform layering, assign adaptive weights to anchor nodes of different layers according to the area where they are located, and propose an adaptive improved Monte Carlo algorithm. After simulation analysis, the algorithm has an average localization error of nodes under different regularity conditions. It has dropped by about 12%, and the localization error has dropped by about 10% on average under different speed conditions. Aiming at the shortcomings of the MCL algorithm in mobile wireless sensor mobile networks such as low localization accuracy, large sample demand, and long localization time, the communication radius of the node is fuzzified to reduce the sampling area of the node, and an improved Monte Carlo localization algorithm based on fuzzy theory is proposed. The improved Monte Carlo localization algorithm, after simulation analysis, is about 50% shorter than the traditional MCL algorithm in localization time, and the localization accuracy is up to about 30% higher than the traditional MCL algorithm.

Wi-Fi indoor 3D localization algorithm based on multi-classifier fusion

Indoor localization technology based on Received Signal Strength (RSS) fingerprint is widely used in life and industry. Compared with the traditional localization methods, the localization technology integrating multiple machine learning methods has better localization accuracy. However, in multi-floor localization, the existing fusion localization technology ignores the interaction between different floors in the localization process, resulting in low indoor localization accuracy. This paper proposes a Wi-Fi indoor 3D localization method based on multi-classifier fusion named FLMCF. Firstly, floor classification training is carried out to reduce the location deviation in the vertical direction. Secondly, for each floor, multiple classifiers are used for model training and training the optimal weight set by minimizing the average localization error. In this case, the advantages of each classifier can be fully integrated to improve the localization accuracy. Finally, the Reliability Fusion Weight Selection (RFWS) algorithm determines the weight and calculates the final location estimation. The experimental results indicate that FLMCF is nearly 14.1% better than DIFMIC in the 90th percentile of CDF.

Bounding‐box deep calibration for high performance face detection

Modern convolutional neural networks (CNNs)-based face detectors have achieved tremendous strides due to large annotated datasets. However, misaligned results with high detection confidence but low localization accuracy restrict the further improvement of detection performance. In this paper, the authors first predict high confidence detection results on the training set itself. Surprisingly, a considerable part of them exist in the same misalignment problem. Then, the authors carefully examine these cases and point out that annotation misalignment is the main reason. Later, a comprehensive discussion is given for the replacement rationality between predicted and annotated bounding-boxes. Finally, the authors propose a novel Bounding-Box Deep Calibration (BDC) method to reasonably replace misaligned annotations with model predicted bounding-boxes and offer calibrated annotations for the training set. Extensive experiments on multiple detectors and two popular benchmark datasets show the effectiveness of BDC on improving models' precision and recall rate, without adding extra inference time and memory consumption. Our simple and effective method provides a general strategy for improving face detection, especially for light-weight detectors in real-time situations.

Node Localization Algorithm Based on Modified Archimedes Optimization Algorithm in Wireless Sensor Networks

Node localization information plays an important role in wireless sensor networks (WSNs). To solve the problem of low localization accuracy of distance vector hop (DV-Hop) localization algorithm in wireless sensor networks, an improved localization algorithm called MAOADV-Hop based on the modified Archimedes optimization algorithm (MAOA) and DV-Hop is proposed, which can achieve the balance between the localization speed and the localization precision. Firstly, tent chaotic mapping and particle swarm optimization (PSO) algorithm are introduced into Archimedes optimization algorithm to improve the initial population diversity and change the update rules of density and volume, which improve the global convergence ability and convergence speed of the algorithm. Secondly, the MAOA is used to replace the least square part of the DV-Hop localization algorithm to improve the localization accuracy of the algorithm. Finally, MAOADV-Hop is verified through four different network environments and compared with DE_DV-Hop, BOA_DV-Hop, and DV-Hop. The simulation results show that the localization speed of the proposed approach is faster than that of DE_DV-Hop and BOA_DV-Hop, and the localization error is less than that of DV-Hop, DE_DV-Hop, and BOA_DV-Hop.

LINKING MULTIPLE PERSPECTIVES WITH OBJECT-BASED VISUAL CUES FOR SPATIAL VIDEO ANALYSIS

Abstract. The visual analysis of videos in context with mapped information requires support in the challenges of linking different spatial perspectives (e.g., street level and survey perspective), bridging different levels of detail, and relating objects in different visual representation. Uncertainty in the spatial relation between camera views and map complicates these tasks. We implemented visualizations for the visual analysis of street level videos (i.e., video key frames) embedded in their spatial context. As part of this, we developed a design rationale for visual cues that help link the video key frames and a map in cases where the spatial relation between camera view and map is of uncertain accuracy. We implemented three cue types (simplified viewshed, object-based dot cues, street centre line cues) for an image data set with heterogeneous camera localization accuracy and assessed the resulting cue properties. Based on this, we argue in favour of cue designs that minimize uncertain information required for their display at the expense of cues’ spatial explicitness in cases of potentially low camera localization accuracy. When localization accuracy is expected to be at least moderate, particularly, dot cues that refer to unambiguous points of reference within easily recognizable objects of ample size present a promising option to support view co-registration.

Distinguishable keypoint detection and matching for optical satellite images with deep convolutional neural networks

As a required step in optical remote sensing applications, image matching identifies correspondences to estimate the relationship between two images. To address this task, feature-based algorithms, such as Scale-Invariant Feature Transform (SIFT), use detectors to identify keypoints and then apply descriptors to represent these keypoints as feature vectors. Thereby, these vectors from different images are matched by Euclidean distance to produce matching points as correspondences. Deep learning networks are widely used in the design of detectors and descriptors due to their powerful representation capabilities. However, previous methods oriented to repeatability tend to be of low localization accuracy. To solve this problem, we define optimal keypoints as “detected, repeatable, reasonable, and distinguishable” (DRRD). Based on this idea, we introduce a DRRD framework for remote sensing optical image matching. Specifically, we designed a deep-learning-based descriptor network which is trained in a self-supervised manner with a novel adaptive mix content triplet (AMCT) loss function. We established a training procedure for remote sensing images using a multi-spectral image dataset. The proposed method obtained 1694 correct correspondences from 8192 detected keypoints in multi-temporal image matching tasks and a sub-pixel accuracy in the image rectification experiments. Comparison experiments show that the proposed DRRD framework outperforms other state-of-the-art methods.

A Novel Range-Free Node Localization Method for Wireless Sensor Networks

Solving the nonconvex and non-differentiable objective function of traditional range-free node localization methods will result in low localization accuracy or high computational complexity. To this end, a novel iterative localization algorithm called CVX-DV-hop is proposed in this letter. It first performs matrix transformation to reformulate the original optimization problem into one with a convex function and nonconvex constraints. And then, the first-order Taylor expansions are employed to tighten the nonconvex constraints into linear inequality constraints. Finally, a successive convex approximation method is designed to iteratively solve the optimization problem. Simulation results show that the proposed algorithm has higher localization accuracy and lower computational burden than those of the particle swarm optimization algorithm.

Transfer learning-based UWB indoor localization using MHT-MDC and clusterization-based sparse fingerprinting

Indoor localization systems have become more and more popular. Several technologies are intensively studied with application to high precision object localization in such environments. Ultra-wideband (UWB) is one of the most promising, as it combines relatively low cost and high localization accuracy, especially compared to Beacon or WiFi. Nevertheless, we noticed that the leading UWB systems’ accuracy is far below the values declared in the documentation. To achieve high localization accuracy, low fingerprinting complexity, and tolerance to significant random errors, we propose a Multiple Hypothesis Tracker with Direction Motion Constraint (MHT-MDC) algorithm followed by a transfer learning approach. We perform fingerprinting using a dense grid in a controlled environment to train the neural network. Thanks to transfer learning, full fingerprinting is not necessary to obtain high localization accuracy when the UWB system is deployed in a new localization. We demonstrate that thanks to transfer learning, high localization accuracy can be maintained when only 7% of fingerprinting samples from a new localization are used to update the neural network, which is very important in practical applications. It is also worth noticing that our approach can be easily extended to other localization technologies. This is an extended version of the conference paper published in [1].

Partial Discharge Pulse Segmentation Approach of Converter Transformers Based on Higher Order Cumulant

As one of the most effective methods to detect the partial discharge (PD) of transformers, high frequency PD detection has been widely used. However, this method also has a bottleneck problem; the biggest problem is the mixed pulse interference under the fixed length sampling. Therefore, this paper focuses on the study of a new pulse segmentation technology, which can separate the partial discharge pulse from the sampling signal containing impulse noise so as to suppress the interference of pulse noise. Based on the characteristics of the high-order-cumulant variation at the rising edge of the pulse signal, a method for judging the starting and ending time of the pulse based on the high-order-cumulant is designed, which can accurately extract the partial discharge pulse from the original data. Simulation results show that the location accuracy of the proposed method can reach 94.67% without stationary noise. The field test shows that the extraction rate of the PD analog signal can reach 79% after applying the segmentation method, which has a great improvement compared with a very low location accuracy rate of 1.65% before using the proposed method.

A novel pipeline leak detection and localization method based on the FBG pipe-fixture sensor array and compressed sensing theory

Low localization accuracy and instability have always been two primary problems for the pipeline leak detection and localization method based on the negative pressure wave (NPW) technique. In this paper, based on compressed sensing (CS) theory, a novel pipeline leak detection and localization method is proposed to replace the solutions of the arrival time difference of NPW knee points. In addition, based on the presented method, the localization accuracy can be improved without increasing the number of sensors. To validate the localization accuracy and stability of the presented method, a pipeline leak simulation test under flow conditions is conducted on a steel pipe leak test platform. The fiber Bragg grating (FBG) pipe-fixture sensor array is developed and applied to detect pipeline leakage and measure the pressure variations in the NPW signals, which provides the data basis for pipeline leak localization based on CS theory. A concept of the relative localization coefficient bCS is proposed, and the test results indicate that when bCS is 0.5, all localization errors of different leakage points with different leakage rates are 2 m.

SOURCE LOCATION MECHANISM OF MICROSEISMIC MONITORING USING P-S WAVES AND ITS EFFECT ANALYSIS

Seismic source location is the most fundamental and most important problem in microseismic monitoring. However, only P wave has been mostly applied in the existing microseismic monitoring networks, with low location accuracy and poor stability of location result for the microseismic events occurring beyond monitoring networks. The seismic source location was implemented using P wave and S wave in this study to expand the effective monitoring area of a microseismic monitoring network and improve its location accuracy for microseismic events nearby the monitoring network. Then, the seismic source location mechanism using P-S wave was revealed through theoretical derivation and analysis. Subsequently, the program development and numerical simulation were combined to analyze and compare systematically the location effects of differently distributed monitoring networks, those consisting of different quantities of sensors, and those with S wave contained in some sensors under two circumstances: combination of P wave and S wave and single use of P wave. Results demonstrate that adding S wave in the plane enhances the accuracy control in the radius direction of the monitoring network. After S wave is included, the location accuracy within a certain area beyond the monitoring network is improved considerably, the effective monitoring area of the whole network is expanded, and the unstable location zones using only P wave are eliminated. The location results of differently distributed monitoring networks and the influence laws of the quantity of sensors constituting the networks on the location results are acquired. This study provides evidence for microseismic monitoring to realize accurate and stable location within a larger range. Keywords: seismic source location, P wave and S wave, mechanism, location effect

Detection of Metal Surface Defects Based on YOLOv4 Algorithm

To solve the problem of low recognition accuracy and low defect location accuracy in traditional detection of surface defects of metal materials, this paper innovates on the basis of YOLOv4 architecture, and studies the influence of adding feature pyramid network module to different position of model neck on detection algorithm. Experiments have shown that adding the feature pyramid network (FPN) module after sampling on the neck network can enhance the feature information expression ability of the feature map originally input to the detection head in the size of 80×80 and 40×40, and achieve better detection results, and achieve better detection results. The experimental results show that adding feature pyramid network module to the neck can effectively improve the detection accuracy of the algorithm. Finally, compared with the traditional YOLOv4 network, the average recognition accuracy of this model can reach 92.5% and the recognition accuracy is improved.

A Closed-Form Localization Method Utilizing Pseudorange Measurements From Two Nonsynchronized Positioning Systems

In a time-of-arrival (TOA) or pseudorange based positioning system, user location is obtained by observing multiple anchor nodes (AN) at known positions. Utilizing more than one positioning systems, e.g., combining Global Positioning System (GPS) and BeiDou Navigation Satellite System (BDS), brings better positioning accuracy. However, ANs from two systems are usually synchronized to two different clock sources. Different from single-system localization, an extra user-to-system clock offset needs to be handled. Existing dual-system methods either have high computational complexity or sub-optimal positioning accuracy. In this paper, we propose a new closed-form dual-system localization (CDL) approach that has low complexity and optimal localization accuracy. We first convert the nonlinear problem into a linear one by squaring the distance equations and employing intermediate variables. Then, a weighted least squares (WLS) method is used to optimize the positioning accuracy. We prove that the positioning error of the new method reaches Cramer-Rao Lower Bound (CRLB) in far field conditions with small measurement noise. Simulations on 2D and 3D positioning scenes are conducted. Results show that, compared with the iterative approach, which has high complexity and requires a good initialization, the new CDL method does not require initialization and has lower computational complexity with comparable positioning accuracy. Numerical results verify the theoretical analysis on positioning accuracy, and show that the new CDL method has superior performance over the state-of-the-art closed-form method. Experiments using real GPS and BDS data verify the applicability of the new CDL method and the superiority of its performance in the real world.

Lightning location method based on improved fuzzy C-means clustering algorithm

Location accuracy is an important index for the evaluation of location networks and the localisation algorithm related to the accuracy of the results. Classical location algorithms scarcely correct errors accurately. Moreover, they have poor resistance to error interference and low location accuracy. To achieve error-resistant lightning localisation, weighted rough-fuzzy C-means (WRFCM) was introduced in location calculation. The performance of this localisation algorithm was analysed using a lightning accident case through regional simulation. Results show that the lightning localisation algorithm based on WFCM overcomes the disadvantage in which traditional location algorithms easily diverge; the algorithm also has an improved ability to resist error interference and can solve the lightning points steadily and accurately.

A number of hops-corrected and hop distance optimization-based DV-Hop localization algorithm

Aiming at the problem of low localization accuracy of classical DV-Hop algorithm, caused by calculation of hop count and hop distance, number of hops-corrected and hop distance optimization-based DV-Hop localization (NHDL) algorithm is proposed in this paper. In NHDL algorithm, the received signal strength values among nodes are used to correct the number of hops. Then, the average distance of each hop of anchor nodes is modified by weighting factor. And the anchor nodes are selected through the concept of collinearity to reduce the localization error. The min-max method is adopted to calculate the position of node. The simulation results show that compared with the traditional DV-Hop algorithm, the average error of localization is reduced.