This paper presents a systematic investigation of the impact of physical parameters such as dimensions, apertures and terminals, on the busbar inductance. A planar laminated busbar with copper conductors and a polyamide insulator is studied using a 3D FEM (Finite Element Modelling) based simulation software. The impact of dimensions such as conductor length, width and thickness, depth of insulator, the effects of aperture number, size, position and connection terminals on the busbar inductance are examined. The magnetic flux density, current density involved and its dependence on the self and mutual inductance of the laminated busbar is analyzed. Laminated busbars provided lower inductance with decreased conductor length, thickness, aperture diameter and depth of insulator. Also, the increased width of the busbar is of the essence for the design of a low inductance busbar. Low inductance laminated busbars are highly beneficial in power converters, powertrain inverters in electric vehicles, photovoltaic converters.
{"title":"Impact of Dimensions, Apertures and Terminals on Stray Inductance of the Laminated Busbar","authors":"Aravind Venugopal, F. Robert","doi":"10.3311/ppee.20827","DOIUrl":"https://doi.org/10.3311/ppee.20827","url":null,"abstract":"This paper presents a systematic investigation of the impact of physical parameters such as dimensions, apertures and terminals, on the busbar inductance. A planar laminated busbar with copper conductors and a polyamide insulator is studied using a 3D FEM (Finite Element Modelling) based simulation software. The impact of dimensions such as conductor length, width and thickness, depth of insulator, the effects of aperture number, size, position and connection terminals on the busbar inductance are examined. The magnetic flux density, current density involved and its dependence on the self and mutual inductance of the laminated busbar is analyzed. Laminated busbars provided lower inductance with decreased conductor length, thickness, aperture diameter and depth of insulator. Also, the increased width of the busbar is of the essence for the design of a low inductance busbar. Low inductance laminated busbars are highly beneficial in power converters, powertrain inverters in electric vehicles, photovoltaic converters.","PeriodicalId":37664,"journal":{"name":"Periodica polytechnica Electrical engineering and computer science","volume":"119 1","pages":"83-94"},"PeriodicalIF":0.0,"publicationDate":"2022-12-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80359147","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
One of the most difficult tasks in the area of computer vision is object detection, which combines object categorization and object location within a scene. In terms of object detection, Deep Neural Networks have been recently demonstrated to outperform alternative approaches. The issues related deep learning neural network is its complexity and huge computation, so it is not possible to detect and track the objects in image of high resolution in real time. We proposed scaled YOLOv4 lite model as Single Stage Detector Neural Network for object detection, tracking and it is trained using COCO 2017 dataset. To create the YOLOv4-CSP- P5- P6- P7- P8 networks, the Scaled YOLOv4 applied efficient network scaling strategies. The additional layer in YOLOv4 lite model is added as P8 layer which improves accuracy. Cross-stage-partial (CSP) connections and Mish activation are used in improved network design, such as backbone optimization and Neck (PAN). In the case of YOLOv4, however, it can only be trained once for all resolutions. Width and Height activations have been changed, allowing for faster network training. With YOLOv4 lite model, we used CSPDarkNet-53 model as a backbone. The experimental result show our YOLOv4 lite model can detect and track object up to 28 fps when model run with the video input and has an accuracy of 86.09% when tested on real-time video with resolutions 1920 × 1080 (full HD). AP = 50.81%, AP @50 = 63.6%, and AP @75 = 52.5% for CSPDarkNet-53 model backbone.
{"title":"Object Classification and Tracking Using Scaled P8 YOLOv4 Lite Model","authors":"Shakil Shaikh, J. Chopade, G. Kharate","doi":"10.3311/ppee.20685","DOIUrl":"https://doi.org/10.3311/ppee.20685","url":null,"abstract":"One of the most difficult tasks in the area of computer vision is object detection, which combines object categorization and object location within a scene. In terms of object detection, Deep Neural Networks have been recently demonstrated to outperform alternative approaches. The issues related deep learning neural network is its complexity and huge computation, so it is not possible to detect and track the objects in image of high resolution in real time. We proposed scaled YOLOv4 lite model as Single Stage Detector Neural Network for object detection, tracking and it is trained using COCO 2017 dataset. To create the YOLOv4-CSP- P5- P6- P7- P8 networks, the Scaled YOLOv4 applied efficient network scaling strategies. The additional layer in YOLOv4 lite model is added as P8 layer which improves accuracy. Cross-stage-partial (CSP) connections and Mish activation are used in improved network design, such as backbone optimization and Neck (PAN). In the case of YOLOv4, however, it can only be trained once for all resolutions. Width and Height activations have been changed, allowing for faster network training. With YOLOv4 lite model, we used CSPDarkNet-53 model as a backbone. The experimental result show our YOLOv4 lite model can detect and track object up to 28 fps when model run with the video input and has an accuracy of 86.09% when tested on real-time video with resolutions 1920 × 1080 (full HD). AP = 50.81%, AP @50 = 63.6%, and AP @75 = 52.5% for CSPDarkNet-53 model backbone.","PeriodicalId":37664,"journal":{"name":"Periodica polytechnica Electrical engineering and computer science","volume":"3 1","pages":"102-111"},"PeriodicalIF":0.0,"publicationDate":"2022-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80120241","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
N. E. H. Yazid, T. Chikouche, K. Hartani, A. Merah
This paper presents an advanced vertical and lateral coupling dynamics control for stabilizing the vertical and lateral motion of a four in-wheel motor drive electric vehicle. The novelty is to investigate the possibilities of enhancing the lateral stability of electric vehicle through the correct control of their suspension system, and further improve the suspension system performances using direct yaw control (DYC) without having a contradictory effect between the vertical and lateral controllers. There are two main objectives; the first is to improve ride comfort, road holding and vehicle lateral handling performance simultaneously by using a new combined full vehicle suspension system, and the second is to strengthen electric vehicle lateral handling performance by using a sliding mode technique for controlling the direct yaw moment. For vehicle suspension system control, we are seeking to design a skyhook control to control the suspension damper and a sliding mode control to stabilize the position of the sprung masses of the in-wheel damper vibration absorber (DVA). To solve the problems caused by the chattering phenomena and ensure the exponential stability of the sprung mass, the continuous singular terminal sliding-mode controller is selected to design the sliding mode control used as a DVA controller. Simulation results show that the coordinated control between DYC and combined full vehicle suspension system under several turns' maneuvers and under random road excitations, can stabilize electric vehicle lateral motion, enhance the vehicle ride comfort, road holding, enhance the rollover resistance performance and satisfy the main suspension performances simultaneously.
{"title":"Advanced Vertical and Lateral Coupling Dynamics Control for a Four In-wheel Motor Drive Electric Vehicle","authors":"N. E. H. Yazid, T. Chikouche, K. Hartani, A. Merah","doi":"10.3311/ppee.20545","DOIUrl":"https://doi.org/10.3311/ppee.20545","url":null,"abstract":"This paper presents an advanced vertical and lateral coupling dynamics control for stabilizing the vertical and lateral motion of a four in-wheel motor drive electric vehicle. The novelty is to investigate the possibilities of enhancing the lateral stability of electric vehicle through the correct control of their suspension system, and further improve the suspension system performances using direct yaw control (DYC) without having a contradictory effect between the vertical and lateral controllers. There are two main objectives; the first is to improve ride comfort, road holding and vehicle lateral handling performance simultaneously by using a new combined full vehicle suspension system, and the second is to strengthen electric vehicle lateral handling performance by using a sliding mode technique for controlling the direct yaw moment. For vehicle suspension system control, we are seeking to design a skyhook control to control the suspension damper and a sliding mode control to stabilize the position of the sprung masses of the in-wheel damper vibration absorber (DVA). To solve the problems caused by the chattering phenomena and ensure the exponential stability of the sprung mass, the continuous singular terminal sliding-mode controller is selected to design the sliding mode control used as a DVA controller. Simulation results show that the coordinated control between DYC and combined full vehicle suspension system under several turns' maneuvers and under random road excitations, can stabilize electric vehicle lateral motion, enhance the vehicle ride comfort, road holding, enhance the rollover resistance performance and satisfy the main suspension performances simultaneously.","PeriodicalId":37664,"journal":{"name":"Periodica polytechnica Electrical engineering and computer science","volume":"17 1","pages":"21-39"},"PeriodicalIF":0.0,"publicationDate":"2022-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78408547","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
An analytical polynomial expression, for accurate and computationally efficient frequency estimation of a single real sinusoid under Additive White Gaussian Noise (AWGN), is derived and proposed in this paper. The method, which can be easily adapted for real time frequency estimation, is based on transforming the frequency estimation problem as the solution of a fourth order (quartic) expressed as powers of a trigonometric function containing the unknown frequency. The coefficients of the quartic polynomial can be found using the complex magnitudes of three Discrete Fourier Transform (DFT) bins, centered at the maximum magnitude value of the DFT coefficients. Simulated results illustrate that, the performance of the proposed estimator has a mean squared-error (MSE) performance which is very close to the Cramer Rao Lower Bound (CRLB) for high signal-to-noise ratio (SNR) region, as well as close to previously published estimators in the low SNR region.
{"title":"Parameter Estimation of a Noisy Real Sinusoid Using Quartic Polynomial Approach","authors":"Gayathri Narayanan, D. Kurup","doi":"10.3311/ppee.20715","DOIUrl":"https://doi.org/10.3311/ppee.20715","url":null,"abstract":"An analytical polynomial expression, for accurate and computationally efficient frequency estimation of a single real sinusoid under Additive White Gaussian Noise (AWGN), is derived and proposed in this paper. The method, which can be easily adapted for real time frequency estimation, is based on transforming the frequency estimation problem as the solution of a fourth order (quartic) expressed as powers of a trigonometric function containing the unknown frequency. The coefficients of the quartic polynomial can be found using the complex magnitudes of three Discrete Fourier Transform (DFT) bins, centered at the maximum magnitude value of the DFT coefficients. Simulated results illustrate that, the performance of the proposed estimator has a mean squared-error (MSE) performance which is very close to the Cramer Rao Lower Bound (CRLB) for high signal-to-noise ratio (SNR) region, as well as close to previously published estimators in the low SNR region.","PeriodicalId":37664,"journal":{"name":"Periodica polytechnica Electrical engineering and computer science","volume":"26 1","pages":"61-69"},"PeriodicalIF":0.0,"publicationDate":"2022-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87736393","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Until now, computational burden alleviation and stability issues for the three-phase four-leg converter has not yet been thoroughly investigated. However, compared to the conventional controllers, the implementation of predictive current control approach for 3-Ø, 4-L inverter suffers a large computational burden due to its additional fourth-leg. Motivated by this fact, this article provides an alternative predictive current control implementation for 3-Ø, 4-L inverter which offers reduced computational effort to achieve similar performance as the conventional FCS-MPC and ensures the global stability of the closed-loop system. To further understand the consequences of the developed control law, theoretical stability analysis has been carried out that links Lyapunov's direct method with the closed-loop system behavior. The outcome of the theoretical stability analysis demonstrates the global stability of the overall system which is later supported by the experimental results. With the proposed method, the number of possible voltage vectors required to obtain the optimal voltage vector in each sampling interval reduces from sixteen to five and thereby simplifies the prediction process. It is also derived that the Lyapunov function-based approach actually yields to the dead-beat control, which has not been previously highlighted in the previous papers. The current work also provides experimental results for different loading conditions (balanced and unbalanced) which further demonstrates the efficacy of the proposed method.
{"title":"Deadbeat Predictive Control Method for 4-leg Inverters","authors":"Tuhin Ibrahim Khan, Hai Tung Luu, L. Számel","doi":"10.3311/ppee.20981","DOIUrl":"https://doi.org/10.3311/ppee.20981","url":null,"abstract":"Until now, computational burden alleviation and stability issues for the three-phase four-leg converter has not yet been thoroughly investigated. However, compared to the conventional controllers, the implementation of predictive current control approach for 3-Ø, 4-L inverter suffers a large computational burden due to its additional fourth-leg. Motivated by this fact, this article provides an alternative predictive current control implementation for 3-Ø, 4-L inverter which offers reduced computational effort to achieve similar performance as the conventional FCS-MPC and ensures the global stability of the closed-loop system. To further understand the consequences of the developed control law, theoretical stability analysis has been carried out that links Lyapunov's direct method with the closed-loop system behavior. The outcome of the theoretical stability analysis demonstrates the global stability of the overall system which is later supported by the experimental results. With the proposed method, the number of possible voltage vectors required to obtain the optimal voltage vector in each sampling interval reduces from sixteen to five and thereby simplifies the prediction process. It is also derived that the Lyapunov function-based approach actually yields to the dead-beat control, which has not been previously highlighted in the previous papers. The current work also provides experimental results for different loading conditions (balanced and unbalanced) which further demonstrates the efficacy of the proposed method.","PeriodicalId":37664,"journal":{"name":"Periodica polytechnica Electrical engineering and computer science","volume":"437 1","pages":"70-82"},"PeriodicalIF":0.0,"publicationDate":"2022-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79616691","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this paper, a Median Grid Search algorithm based on the median values is proposed to develop a more accurate algorithm used to investigate the optimal hyperparameter of the SVR model for load forecasting. In addition, the methodology to benchmark the proposed Grid Search and the conventional Grid Search is built and sufficiently utilized. The data gathered from the South Australia state, Australia and Ho Chi Minh City load demands are used in experiments. Experimental results demonstrate that the proposed algorithm outperforms the conventional algorithm.
{"title":"A New Grid Search Algorithm Based on Median Values for SVR Model in Case of Load Forecasting","authors":"ThanhNgoc Tran","doi":"10.3311/ppee.20887","DOIUrl":"https://doi.org/10.3311/ppee.20887","url":null,"abstract":"In this paper, a Median Grid Search algorithm based on the median values is proposed to develop a more accurate algorithm used to investigate the optimal hyperparameter of the SVR model for load forecasting. In addition, the methodology to benchmark the proposed Grid Search and the conventional Grid Search is built and sufficiently utilized. The data gathered from the South Australia state, Australia and Ho Chi Minh City load demands are used in experiments. Experimental results demonstrate that the proposed algorithm outperforms the conventional algorithm.","PeriodicalId":37664,"journal":{"name":"Periodica polytechnica Electrical engineering and computer science","volume":"50 1","pages":"51-60"},"PeriodicalIF":0.0,"publicationDate":"2022-10-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"72530600","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The Direct Torque Control (DTC) technique requires stator currents and DC bus voltage, as well as inverter switch states, in order to estimate stator flux and electromagnetic torque values. Measurement of phase currents in real time with current sensors is a common method of gathering this data. To control a Five-Phase Induction Motor (FPIM), the DTC needs at least four alternating current (AC) sensors and one voltage sensor. When utilized in a global training system, this number has disadvantages due to cost, size, and non-linearity. The purpose of this article is to show how to remove the alternating current sensors from an FPIM's DTC. This is accomplished by reconstructing the phase currents using a simple DC current sensor placed at the voltage inverter's input and modifying the classical DTC control technique by using a good choice of certain allowable switching vectors of a five-phase voltage inverter, the 18° zone offset strategy of the stator flux, and developing a new commutation table for the stator currents construction. The proposed control approach is supported by simulation results.
{"title":"A Highly Effective Modified Direct Torque Control for Five Phase Induction Motor without AC Phase Current Sensors","authors":"A. Azib, D. Ziane","doi":"10.3311/ppee.20384","DOIUrl":"https://doi.org/10.3311/ppee.20384","url":null,"abstract":"The Direct Torque Control (DTC) technique requires stator currents and DC bus voltage, as well as inverter switch states, in order to estimate stator flux and electromagnetic torque values. Measurement of phase currents in real time with current sensors is a common method of gathering this data. To control a Five-Phase Induction Motor (FPIM), the DTC needs at least four alternating current (AC) sensors and one voltage sensor. When utilized in a global training system, this number has disadvantages due to cost, size, and non-linearity. The purpose of this article is to show how to remove the alternating current sensors from an FPIM's DTC. This is accomplished by reconstructing the phase currents using a simple DC current sensor placed at the voltage inverter's input and modifying the classical DTC control technique by using a good choice of certain allowable switching vectors of a five-phase voltage inverter, the 18° zone offset strategy of the stator flux, and developing a new commutation table for the stator currents construction. The proposed control approach is supported by simulation results.","PeriodicalId":37664,"journal":{"name":"Periodica polytechnica Electrical engineering and computer science","volume":"36 1","pages":"1-9"},"PeriodicalIF":0.0,"publicationDate":"2022-10-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81112477","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In a recent paper, the conventional sidelobe canceller radar system was developed by replacing the separate auxiliary antennas by few elements at the center of the main antenna array. The modified system with reused elements was associated with some attenuation in the desired signal due to the emerging correlation between the signals that exists in the main and the reused array elements. This problem was solved by imposing some constraints on the array pattern of the reused elements. In this paper, few of the side elements of the main array are employed as the auxiliary antennas. This new proposed configuration is called sided-elements. Unlike the previous centered-elements configuration, the proposed sided-elements configuration offers more desired features since the pattern of the side elements has sidelobes of similar widths of those of the main array. Moreover, a better diversity is obtained due to the wider separation between the two groups of elements at both sides of the main array. Simulation results fully confirm the effectiveness of the new proposed sided-elements configuration for suppressing the undesired interfering signals and retaining the desired signal undistorted.
{"title":"Elements Selection for Auxiliary Array in the Adaptive Sidelobe Canceller Radar System","authors":"J. Mohammed","doi":"10.3311/ppee.20452","DOIUrl":"https://doi.org/10.3311/ppee.20452","url":null,"abstract":"In a recent paper, the conventional sidelobe canceller radar system was developed by replacing the separate auxiliary antennas by few elements at the center of the main antenna array. The modified system with reused elements was associated with some attenuation in the desired signal due to the emerging correlation between the signals that exists in the main and the reused array elements. This problem was solved by imposing some constraints on the array pattern of the reused elements. In this paper, few of the side elements of the main array are employed as the auxiliary antennas. This new proposed configuration is called sided-elements. Unlike the previous centered-elements configuration, the proposed sided-elements configuration offers more desired features since the pattern of the side elements has sidelobes of similar widths of those of the main array. Moreover, a better diversity is obtained due to the wider separation between the two groups of elements at both sides of the main array. Simulation results fully confirm the effectiveness of the new proposed sided-elements configuration for suppressing the undesired interfering signals and retaining the desired signal undistorted.","PeriodicalId":37664,"journal":{"name":"Periodica polytechnica Electrical engineering and computer science","volume":"84 1","pages":"406-414"},"PeriodicalIF":0.0,"publicationDate":"2022-10-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81593139","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Fatima Belgacem, M. Mostefai, Miloud Yahia, Aicha Belgacem
This study presents a comparative analysis of different MPPT techniques for PV array fed water pumping system without batteries using brushless direct current (BLDC) Motor. The system is configured to reduce both cost and complexity with simultaneous assurance of optimum power utilization of PV array, for this we propose two intelligent maximum power point tracking MPPT techniques implemented through a boost converter; the first is based on Fuzzy Logic Controller (FLC) the second is a sliding mode control, which have been successfully applied to Photovoltaic Pumping System (PVPS) to replace tracking power point perturb and observe (P&O) under different conditions of solar radiation. As a result, the PV must be turned on at its highest power extracted. The water pumping system is designed to irrigate an area of 8 hectares, planted with onions, we have chosen a BLDC motor which is coupled with a centrifugal pump to overcome the disadvantage of brushed motors and induction motors. The object is to verify and confirm the effectiveness of the suggested system. According to simulation data, FLC improves the PVPS's performance.
{"title":"Optimization of Photovoltaic Water Pumping System Based on BLDC Motor for Agricultural Irrigation with Different MPPT Methods","authors":"Fatima Belgacem, M. Mostefai, Miloud Yahia, Aicha Belgacem","doi":"10.3311/ppee.20140","DOIUrl":"https://doi.org/10.3311/ppee.20140","url":null,"abstract":"This study presents a comparative analysis of different MPPT techniques for PV array fed water pumping system without batteries using brushless direct current (BLDC) Motor. The system is configured to reduce both cost and complexity with simultaneous assurance of optimum power utilization of PV array, for this we propose two intelligent maximum power point tracking MPPT techniques implemented through a boost converter; the first is based on Fuzzy Logic Controller (FLC) the second is a sliding mode control, which have been successfully applied to Photovoltaic Pumping System (PVPS) to replace tracking power point perturb and observe (P&O) under different conditions of solar radiation. As a result, the PV must be turned on at its highest power extracted. The water pumping system is designed to irrigate an area of 8 hectares, planted with onions, we have chosen a BLDC motor which is coupled with a centrifugal pump to overcome the disadvantage of brushed motors and induction motors. The object is to verify and confirm the effectiveness of the suggested system. According to simulation data, FLC improves the PVPS's performance.","PeriodicalId":37664,"journal":{"name":"Periodica polytechnica Electrical engineering and computer science","volume":"32 1","pages":"315-324"},"PeriodicalIF":0.0,"publicationDate":"2022-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"83102459","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This paper presents a new approach for the challenging problem of image geo-localization using Convolutional Neural Networks (CNNs). This latter has become the state-of-the-art technique in computer vision and machine learning, particularly in location recognition of images taken in urban environments where the recognition accuracy is very impressive. We cast this task as a classification problem. First, we extract features from images by using pre-trained CNN model AlexNet as a feature extraction tool; where the output of the fully connected layer is considered as the feature representation. Then, the features extracted from the fully connected layer can be used for the classification process by feeding them into the Support Vector Machine (SVM) classifier. We evaluated the proposed approach on a data set of Google Street View images (GSV); the experimental results show that our approach can improve the classification by achieving a good accuracy rate which is 94.19%.
{"title":"Improving Street View Image Classification Using Pre-trained CNN Model Extracted Features","authors":"Meriem Djouadi, M. Kholladi","doi":"10.3311/ppee.19961","DOIUrl":"https://doi.org/10.3311/ppee.19961","url":null,"abstract":"This paper presents a new approach for the challenging problem of image geo-localization using Convolutional Neural Networks (CNNs). This latter has become the state-of-the-art technique in computer vision and machine learning, particularly in location recognition of images taken in urban environments where the recognition accuracy is very impressive. We cast this task as a classification problem. First, we extract features from images by using pre-trained CNN model AlexNet as a feature extraction tool; where the output of the fully connected layer is considered as the feature representation. Then, the features extracted from the fully connected layer can be used for the classification process by feeding them into the Support Vector Machine (SVM) classifier. We evaluated the proposed approach on a data set of Google Street View images (GSV); the experimental results show that our approach can improve the classification by achieving a good accuracy rate which is 94.19%.","PeriodicalId":37664,"journal":{"name":"Periodica polytechnica Electrical engineering and computer science","volume":"1 1","pages":"370-379"},"PeriodicalIF":0.0,"publicationDate":"2022-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"76481491","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：