Pub Date : 2022-11-01DOI: 10.1109/CCISP55629.2022.9974564
R. Shi, Sanxin Jiang
It is a necessary task to extract and save structured information from invoices. The existing methods are all to detect and identify the duplication of invoices. Considering that there are a lot of duplicate contents and fixed table structure between invoices of the same type, this method proposes to separate the template and filled contents of invoices by pixel segmentation; The perceptual hash algorithm is used to match the template of the invoice to be tested with the invoice in the template database; After successful matching, use the improved template alignment module to align the new filled content with the template invoice, and then import the new invoice into Excel for saving. Experimental results show that compared with the original method, the text detection time, recognition time and prediction time of this method are reduced by 68%, 91.13% and 89.94% respectively, and the overall prediction time is reduced by 27.26 seconds.
{"title":"Invoice Recognition Method Based on Separation of Template and Content","authors":"R. Shi, Sanxin Jiang","doi":"10.1109/CCISP55629.2022.9974564","DOIUrl":"https://doi.org/10.1109/CCISP55629.2022.9974564","url":null,"abstract":"It is a necessary task to extract and save structured information from invoices. The existing methods are all to detect and identify the duplication of invoices. Considering that there are a lot of duplicate contents and fixed table structure between invoices of the same type, this method proposes to separate the template and filled contents of invoices by pixel segmentation; The perceptual hash algorithm is used to match the template of the invoice to be tested with the invoice in the template database; After successful matching, use the improved template alignment module to align the new filled content with the template invoice, and then import the new invoice into Excel for saving. Experimental results show that compared with the original method, the text detection time, recognition time and prediction time of this method are reduced by 68%, 91.13% and 89.94% respectively, and the overall prediction time is reduced by 27.26 seconds.","PeriodicalId":431851,"journal":{"name":"2022 7th International Conference on Communication, Image and Signal Processing (CCISP)","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134212534","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}
Pub Date : 2022-11-01DOI: 10.1109/ccisp55629.2022.9974168
{"title":"Sponsored","authors":"","doi":"10.1109/ccisp55629.2022.9974168","DOIUrl":"https://doi.org/10.1109/ccisp55629.2022.9974168","url":null,"abstract":"","PeriodicalId":431851,"journal":{"name":"2022 7th International Conference on Communication, Image and Signal Processing (CCISP)","volume":"102 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131058330","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}
Pub Date : 2022-11-01DOI: 10.1109/CCISP55629.2022.9974477
Liang Futai, Zhou Yan, Zhang Chenhao, Song Zihao, Zhao Xiaorui
Aircraft formation recognition is of great significance in the intention prediction and the threat assessment field, but the current traditional template-based methods need to manually extract features and construct templates, which has the problems of complex process and poor effect. This paper proposes a formation recognition method based on GAN and CNN, which can perform end-to-end formation recognition. First, a GAN model is designed to generate a large amount of new aircraft formation data from a small amount of measured data. Then a CNN-based aircraft formation recognition model is designed. After the model training is completed, the aircraft formation recognition can be completed by inputting the measured aircraft formation data. The experimental results show that this method can improve the recognition accuracy by 8%.
{"title":"An aircraft formation recognition method based on deep learning","authors":"Liang Futai, Zhou Yan, Zhang Chenhao, Song Zihao, Zhao Xiaorui","doi":"10.1109/CCISP55629.2022.9974477","DOIUrl":"https://doi.org/10.1109/CCISP55629.2022.9974477","url":null,"abstract":"Aircraft formation recognition is of great significance in the intention prediction and the threat assessment field, but the current traditional template-based methods need to manually extract features and construct templates, which has the problems of complex process and poor effect. This paper proposes a formation recognition method based on GAN and CNN, which can perform end-to-end formation recognition. First, a GAN model is designed to generate a large amount of new aircraft formation data from a small amount of measured data. Then a CNN-based aircraft formation recognition model is designed. After the model training is completed, the aircraft formation recognition can be completed by inputting the measured aircraft formation data. The experimental results show that this method can improve the recognition accuracy by 8%.","PeriodicalId":431851,"journal":{"name":"2022 7th International Conference on Communication, Image and Signal Processing (CCISP)","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132984410","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}
Pub Date : 2022-11-01DOI: 10.1109/CCISP55629.2022.9974264
Xin Yu, Xingyuan You, Xiaoling Liu, Chuanjun Li
Under narrowband shortwave communication conditions, digital speech coding is mostly in the form of low-rate linear predictive coding, but LPC parametric coding recovers low naturalness of speech with buzz. In this paper, we propose a method to improve the residual signal excitation based on LPC10. At the coding end, the prediction coefficients are solved based on linear prediction analysis, and the original speech is inverse filtered based on the prediction coefficients and differs from the original speech signal to obtain the residual signal; at the decoding end, the original muffled pulse excitation is replaced with the residual signal, and the improved synthesized speech improves the hum in the original LPC synthesized speech. The generated speech and the original speech are scored by PESQ algorithm, and the result showed that the improved speech score is 1.68, which is 0.34 points higher than the LPC 10 synthesized speech score.
{"title":"An improved algorithm for residual signal excitation based on LPC 10","authors":"Xin Yu, Xingyuan You, Xiaoling Liu, Chuanjun Li","doi":"10.1109/CCISP55629.2022.9974264","DOIUrl":"https://doi.org/10.1109/CCISP55629.2022.9974264","url":null,"abstract":"Under narrowband shortwave communication conditions, digital speech coding is mostly in the form of low-rate linear predictive coding, but LPC parametric coding recovers low naturalness of speech with buzz. In this paper, we propose a method to improve the residual signal excitation based on LPC10. At the coding end, the prediction coefficients are solved based on linear prediction analysis, and the original speech is inverse filtered based on the prediction coefficients and differs from the original speech signal to obtain the residual signal; at the decoding end, the original muffled pulse excitation is replaced with the residual signal, and the improved synthesized speech improves the hum in the original LPC synthesized speech. The generated speech and the original speech are scored by PESQ algorithm, and the result showed that the improved speech score is 1.68, which is 0.34 points higher than the LPC 10 synthesized speech score.","PeriodicalId":431851,"journal":{"name":"2022 7th International Conference on Communication, Image and Signal Processing (CCISP)","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133658729","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}
Pub Date : 2022-11-01DOI: 10.1109/CCISP55629.2022.9974202
Changzhi Li, Xi Yang, Hang Zhang, Linyuan Wang, Jiayi Wan
Given the inconvenience of air conditioning control methods based on control panels or remote controls, the application of an improved K-Nearest Neighbor (KNN) algorithm gesture recognition system based on the entropy weight method in air conditioning control is carried out. First, establish the correspondence between the ten commonly used gestures from 1 to 10 and the air conditioning control commands, then collect the gesture images from 1 to 10 through the camera, and then perform image preprocessing, gesture contour extraction, and feature calculation. In the Euclidean distance calculation process, the weight coefficient determined by the entropy weight method is added, and the trained improved KNN model is used to recognize the gesture, thereby improving the accuracy of the gesture recognition process. Simulation studies show that the accuracy of the gesture recognition system based on the improved KNN model is over 95%. This result is 9.7%-11.6% higher than that of the conventional KNN model before improvement and 11.8%-12.9% higher than the accuracy of the support vector machine algorithm (SVM) model. The experimental results show that the accuracy of the gesture recognition system based on the improved KNN algorithm is between 77.5% and 87.5%. Therefore, the method proposed in this paper has a good application prospect in air-conditioning control.
{"title":"Application of improved K-Nearest Neighbor algorithm gesture recognition system in air conditioning control","authors":"Changzhi Li, Xi Yang, Hang Zhang, Linyuan Wang, Jiayi Wan","doi":"10.1109/CCISP55629.2022.9974202","DOIUrl":"https://doi.org/10.1109/CCISP55629.2022.9974202","url":null,"abstract":"Given the inconvenience of air conditioning control methods based on control panels or remote controls, the application of an improved K-Nearest Neighbor (KNN) algorithm gesture recognition system based on the entropy weight method in air conditioning control is carried out. First, establish the correspondence between the ten commonly used gestures from 1 to 10 and the air conditioning control commands, then collect the gesture images from 1 to 10 through the camera, and then perform image preprocessing, gesture contour extraction, and feature calculation. In the Euclidean distance calculation process, the weight coefficient determined by the entropy weight method is added, and the trained improved KNN model is used to recognize the gesture, thereby improving the accuracy of the gesture recognition process. Simulation studies show that the accuracy of the gesture recognition system based on the improved KNN model is over 95%. This result is 9.7%-11.6% higher than that of the conventional KNN model before improvement and 11.8%-12.9% higher than the accuracy of the support vector machine algorithm (SVM) model. The experimental results show that the accuracy of the gesture recognition system based on the improved KNN algorithm is between 77.5% and 87.5%. Therefore, the method proposed in this paper has a good application prospect in air-conditioning control.","PeriodicalId":431851,"journal":{"name":"2022 7th International Conference on Communication, Image and Signal Processing (CCISP)","volume":"87 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131308896","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}
Optical fiber communication has been rapidly developed and applied because of its huge transmission capacity and minimumness transmission attenuation. However, due to the difference in frequency or mode of the transmitted optical signals, dispersion is very likely to occur, causing overlap between pulses, which leads to inter-code interference and affects the judgment of the adjudicator at the receiving end, thus reducing the capacity of communication. In this paper, a high-precision dispersion measurement system based on the clock phase shift method is designed to improve the measurement accuracy and reduce the cost of measurement. The design uses an XILINX's Spartan®-7 series FPGA (Field Programmable Gate Array) chip to achieve accurate measurement of phase difference between counting clocks. The overall modular design idea, using logic control, realized the data communication between modules and register read/write control, and completed the data acquisition and processing. The performance of the system is experimentally verified, and the experimental results show that the measurement error can reach ±500ps for the long-distance fiber of 81km.
{"title":"Optical fiber dispersion measurement system based on frequency phase shift and its FPGA implementation","authors":"Yuling Wang, X. Han, Cheng Liu, Qian Wang, Zheliang Zhang","doi":"10.1109/CCISP55629.2022.9974529","DOIUrl":"https://doi.org/10.1109/CCISP55629.2022.9974529","url":null,"abstract":"Optical fiber communication has been rapidly developed and applied because of its huge transmission capacity and minimumness transmission attenuation. However, due to the difference in frequency or mode of the transmitted optical signals, dispersion is very likely to occur, causing overlap between pulses, which leads to inter-code interference and affects the judgment of the adjudicator at the receiving end, thus reducing the capacity of communication. In this paper, a high-precision dispersion measurement system based on the clock phase shift method is designed to improve the measurement accuracy and reduce the cost of measurement. The design uses an XILINX's Spartan®-7 series FPGA (Field Programmable Gate Array) chip to achieve accurate measurement of phase difference between counting clocks. The overall modular design idea, using logic control, realized the data communication between modules and register read/write control, and completed the data acquisition and processing. The performance of the system is experimentally verified, and the experimental results show that the measurement error can reach ±500ps for the long-distance fiber of 81km.","PeriodicalId":431851,"journal":{"name":"2022 7th International Conference on Communication, Image and Signal Processing (CCISP)","volume":"231 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115907472","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}
Pub Date : 2022-11-01DOI: 10.1109/CCISP55629.2022.9974277
Jiajun Wu, Jun Gao, Jiangang Yi, P. Liu, Changsong Xu
Environmental perception is a necessary prerequisite for intelligent robots to perform specified tasks, and is the basis for subsequent control and decision-making. In recent years, with the rapid development of deep learning technology and the dramatic improvement of hardware performance, vision-based environmental perception technologies, such as target recognition and target detection, have made significant progress. However, most vision algorithms are developed based on images with stable lighting conditions and no significant disturbances. In fact, robots often need to operate in unstructured, complex conditions or visually degraded environments. Visual perception alone cannot meet the job requirements and it lacks the ability to adapt to the environment. Therefore, the environment perception technology based on multi-sensor fusion has become a popular research direction. In this paper, we first analyze the characteristics of sensors required for perception, and briefly review the uni-modal sensor application status in complex environments such as mines, railways, highways, tunnels, etc. Secondly, we introduce the datasets and sensor fusion methods for robotics perception. Thirdly, we provide an overview of the multi-modal perception technology applied on intelligent robot. Finally, we summarize the challenges and future development trends in this direction.
{"title":"Environment Perception Technology for Intelligent Robots in Complex Environments: A Review","authors":"Jiajun Wu, Jun Gao, Jiangang Yi, P. Liu, Changsong Xu","doi":"10.1109/CCISP55629.2022.9974277","DOIUrl":"https://doi.org/10.1109/CCISP55629.2022.9974277","url":null,"abstract":"Environmental perception is a necessary prerequisite for intelligent robots to perform specified tasks, and is the basis for subsequent control and decision-making. In recent years, with the rapid development of deep learning technology and the dramatic improvement of hardware performance, vision-based environmental perception technologies, such as target recognition and target detection, have made significant progress. However, most vision algorithms are developed based on images with stable lighting conditions and no significant disturbances. In fact, robots often need to operate in unstructured, complex conditions or visually degraded environments. Visual perception alone cannot meet the job requirements and it lacks the ability to adapt to the environment. Therefore, the environment perception technology based on multi-sensor fusion has become a popular research direction. In this paper, we first analyze the characteristics of sensors required for perception, and briefly review the uni-modal sensor application status in complex environments such as mines, railways, highways, tunnels, etc. Secondly, we introduce the datasets and sensor fusion methods for robotics perception. Thirdly, we provide an overview of the multi-modal perception technology applied on intelligent robot. Finally, we summarize the challenges and future development trends in this direction.","PeriodicalId":431851,"journal":{"name":"2022 7th International Conference on Communication, Image and Signal Processing (CCISP)","volume":"58 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124177690","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}
Pub Date : 2022-11-01DOI: 10.1109/CCISP55629.2022.9974568
Xin Tan, Zhidong Xie, Xinwang Yuan, Gang Yang, Yung-Su Han
It is difficult and costly to obtain sufficient label samples in an open environment, and the study of small sample problem has become an important direction in the field of signal modulation recognition. For the first time, we innovatively propose to use CatBoost to solve this. First, we extract high-order cumulants features from the I/Q signal, and then add slicing operations to make this feature suitable for algorithm training, and finally take advantage of CatBoost's high classification accuracy under small sample condition, to achieve effective recognition of small sample signals. The experiment obtained the results of comprehensive recognition accuracy of 93.3% and 95.1% when there are 20 and 200 samples in each type of 9 types of signals from 0 to 8dB, respectively. Compared with other traditional machine learning algorithms and deep learning algorithms, our method is more efficient.
{"title":"Small Sample Signal Modulation Recognition based on Higher-order Cumulants and CatBoost","authors":"Xin Tan, Zhidong Xie, Xinwang Yuan, Gang Yang, Yung-Su Han","doi":"10.1109/CCISP55629.2022.9974568","DOIUrl":"https://doi.org/10.1109/CCISP55629.2022.9974568","url":null,"abstract":"It is difficult and costly to obtain sufficient label samples in an open environment, and the study of small sample problem has become an important direction in the field of signal modulation recognition. For the first time, we innovatively propose to use CatBoost to solve this. First, we extract high-order cumulants features from the I/Q signal, and then add slicing operations to make this feature suitable for algorithm training, and finally take advantage of CatBoost's high classification accuracy under small sample condition, to achieve effective recognition of small sample signals. The experiment obtained the results of comprehensive recognition accuracy of 93.3% and 95.1% when there are 20 and 200 samples in each type of 9 types of signals from 0 to 8dB, respectively. Compared with other traditional machine learning algorithms and deep learning algorithms, our method is more efficient.","PeriodicalId":431851,"journal":{"name":"2022 7th International Conference on Communication, Image and Signal Processing (CCISP)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114668842","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}
Pub Date : 2022-11-01DOI: 10.1109/CCISP55629.2022.9974370
Xin Shang, R. Ye, Hui-Na Feng, Xueqin Jiang
Point cloud is an important and commonly used signal representation for volume objects or scenes in the real world. Due to the imperfect acquisition of the point cloud, there is nonnegligible noise in the point cloud. Most literatures that use graph signal processing (GSP) for point cloud denoising (PCD) generally construct k-NN graph to represent the point cloud. However, the graph constructed based on this scheme can not compactly represent the underlying structure of a noisy point cloud. In this paper, we propose a feature graph that can effectively and naturally represent the structure of the point cloud. To construct the feature graph, a feature sampling method is exploited to obtain the feature points. Then, patches are built based on the feature points. After that, the feature graph is constructed by connecting all the points in the patches. Finally, we apply the feature graph to the PCD problem and exploit graph Laplacian regularization (GLR) as smoothing prior information for denoising. Experimental results show that our proposed PCD method not only outperforms the existing PCD methods in objective evaluation metrics, but also performs better in processing the inner and edge structure of the point cloud.
{"title":"Robust Feature Graph for Point Cloud Denoising","authors":"Xin Shang, R. Ye, Hui-Na Feng, Xueqin Jiang","doi":"10.1109/CCISP55629.2022.9974370","DOIUrl":"https://doi.org/10.1109/CCISP55629.2022.9974370","url":null,"abstract":"Point cloud is an important and commonly used signal representation for volume objects or scenes in the real world. Due to the imperfect acquisition of the point cloud, there is nonnegligible noise in the point cloud. Most literatures that use graph signal processing (GSP) for point cloud denoising (PCD) generally construct k-NN graph to represent the point cloud. However, the graph constructed based on this scheme can not compactly represent the underlying structure of a noisy point cloud. In this paper, we propose a feature graph that can effectively and naturally represent the structure of the point cloud. To construct the feature graph, a feature sampling method is exploited to obtain the feature points. Then, patches are built based on the feature points. After that, the feature graph is constructed by connecting all the points in the patches. Finally, we apply the feature graph to the PCD problem and exploit graph Laplacian regularization (GLR) as smoothing prior information for denoising. Experimental results show that our proposed PCD method not only outperforms the existing PCD methods in objective evaluation metrics, but also performs better in processing the inner and edge structure of the point cloud.","PeriodicalId":431851,"journal":{"name":"2022 7th International Conference on Communication, Image and Signal Processing (CCISP)","volume":"30 6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125701491","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}
Pub Date : 2022-11-01DOI: 10.1109/CCISP55629.2022.9974601
Zongxi Zhang, Mingfu Fu, Jie Mei, Ming Zhu, Jing Zhang, Lingjun Xiao
High voltage shunt reactor is a primary electric power apparatus and plays a significance role in electric power system. In term of diagnosing the fault of high voltage shunt reactor, vibration signal is an easy acquired information. However, in the initial fault stage of high voltage shunt reactor, the characteristic information of vibration signal is weak and the noise interference is large. In this paper, we collect vibration signal from 24 sampling position on the four sides of high voltage shunt reactor under different kinds of state. We put forward a way which is based on spectral residual and machine learning to diagnosis high voltage shunt reactors fault. This method not only can effectively remove the weak direct current component in the frequency component but also can highlight the fundamental frequency component and frequency doubling component. In the experiment, we set up the fault diagnosis models by Support Vector Machine (SVM) and Convolutional Neural Network (CNN) respectively to compare the residual signals of raw vibration signal spectrum. The results show that compared to the raw vibration signal, the spectrum residual algorithm improved accuracy state by 9% and 10.75% respectively. Therefore, spectral residual can improve the fault diagnosis accuracy of high voltage shunt reactors.
{"title":"A new machine learning-basd fault diagnosis method of high voltage shunt reactor using spectral residual","authors":"Zongxi Zhang, Mingfu Fu, Jie Mei, Ming Zhu, Jing Zhang, Lingjun Xiao","doi":"10.1109/CCISP55629.2022.9974601","DOIUrl":"https://doi.org/10.1109/CCISP55629.2022.9974601","url":null,"abstract":"High voltage shunt reactor is a primary electric power apparatus and plays a significance role in electric power system. In term of diagnosing the fault of high voltage shunt reactor, vibration signal is an easy acquired information. However, in the initial fault stage of high voltage shunt reactor, the characteristic information of vibration signal is weak and the noise interference is large. In this paper, we collect vibration signal from 24 sampling position on the four sides of high voltage shunt reactor under different kinds of state. We put forward a way which is based on spectral residual and machine learning to diagnosis high voltage shunt reactors fault. This method not only can effectively remove the weak direct current component in the frequency component but also can highlight the fundamental frequency component and frequency doubling component. In the experiment, we set up the fault diagnosis models by Support Vector Machine (SVM) and Convolutional Neural Network (CNN) respectively to compare the residual signals of raw vibration signal spectrum. The results show that compared to the raw vibration signal, the spectrum residual algorithm improved accuracy state by 9% and 10.75% respectively. Therefore, spectral residual can improve the fault diagnosis accuracy of high voltage shunt reactors.","PeriodicalId":431851,"journal":{"name":"2022 7th International Conference on Communication, Image and Signal Processing (CCISP)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126127744","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：