Pub Date : 2022-07-08DOI: 10.1109/ICMSP55950.2022.9859077
Yangyang Jia, Shourui Yang
In fringe projection measurement, the suitable defocus of the projector can improve the measurement accuracy. This paper presents a method to determine the optimal defocus level in different object distances. In each object distance, this method adjusts the focal length of the projector through some special patterns, and the plane plate is measured by binocular fringe projection measurement system in different defocus levels. Then, the fitting plane error in different defocus levels is polynomial fitted, and the optimal defocus degrees are calculated on the fitting curve. Finally, the corresponding relationship between the object distance and the optimal defocus level is obtained by spline interpolation fitting. Experimental results show that this method can be successfully applied to fringe projection measurement with any object distance, and the measurement accuracy is improved by more than 10%.
{"title":"Determination of the relationship between object distance and optimal defocus level for fringe projection systems","authors":"Yangyang Jia, Shourui Yang","doi":"10.1109/ICMSP55950.2022.9859077","DOIUrl":"https://doi.org/10.1109/ICMSP55950.2022.9859077","url":null,"abstract":"In fringe projection measurement, the suitable defocus of the projector can improve the measurement accuracy. This paper presents a method to determine the optimal defocus level in different object distances. In each object distance, this method adjusts the focal length of the projector through some special patterns, and the plane plate is measured by binocular fringe projection measurement system in different defocus levels. Then, the fitting plane error in different defocus levels is polynomial fitted, and the optimal defocus degrees are calculated on the fitting curve. Finally, the corresponding relationship between the object distance and the optimal defocus level is obtained by spline interpolation fitting. Experimental results show that this method can be successfully applied to fringe projection measurement with any object distance, and the measurement accuracy is improved by more than 10%.","PeriodicalId":114259,"journal":{"name":"2022 4th International Conference on Intelligent Control, Measurement and Signal Processing (ICMSP)","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133937199","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 : 2022-07-08DOI: 10.1109/ICMSP55950.2022.9859035
Shuaiyin Zhang, Zhongyu Wang, Yinbao Cheng, Xuejiao Dai
This paper describes an evaluation method of free-form curve profile, which uses the distribution characteristics of directed profile error. This method can express the deviation trend of curve contour and the quality of overall contour to a certain extent. It adopts the concept of directed profile error (DPE), which can quantify the feature of the difference between the ideal curve and the actual curve, including size and direction of deviation. To eliminate systematic errors, the calculation of DPE mainly includes two steps, curve registration and curve fitting. Aiming to evaluate the free-form curve profile, this paper analyzes the degree of concentration and dispersion of DPE. The simulation experiment and actual measurement experiment is designed to test the effectiveness and applicability of this method. From the experimental results, this method can be used to describe the profile features of free-form curves better and meet the needs of quality assessment.
{"title":"An evaluation method for free- form curve profile based on the distribution characteristics of directed profile error","authors":"Shuaiyin Zhang, Zhongyu Wang, Yinbao Cheng, Xuejiao Dai","doi":"10.1109/ICMSP55950.2022.9859035","DOIUrl":"https://doi.org/10.1109/ICMSP55950.2022.9859035","url":null,"abstract":"This paper describes an evaluation method of free-form curve profile, which uses the distribution characteristics of directed profile error. This method can express the deviation trend of curve contour and the quality of overall contour to a certain extent. It adopts the concept of directed profile error (DPE), which can quantify the feature of the difference between the ideal curve and the actual curve, including size and direction of deviation. To eliminate systematic errors, the calculation of DPE mainly includes two steps, curve registration and curve fitting. Aiming to evaluate the free-form curve profile, this paper analyzes the degree of concentration and dispersion of DPE. The simulation experiment and actual measurement experiment is designed to test the effectiveness and applicability of this method. From the experimental results, this method can be used to describe the profile features of free-form curves better and meet the needs of quality assessment.","PeriodicalId":114259,"journal":{"name":"2022 4th International Conference on Intelligent Control, Measurement and Signal Processing (ICMSP)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134356035","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 : 2022-07-08DOI: 10.1109/ICMSP55950.2022.9859113
Jingrui Chu, Shi-kui Wang, Gang Yang, Wenhang Wang
Aiming at the simulation requirements of airflow temperature signal for the total temperature simulation device of hypersonic aircraft, an airflow temperature simulation system is designed. The system consists of a high-temperature plasma spray gun, an airflow mixing chamber, and a thermocouple. The maximal temperature of the system is high and the temperature changes rapidly. But the thermocouple responds slowly, hence there is a large error when measuring the airflow temperature. However, the accuracy of airflow temperature measurement results is important to improve the accuracy of airflow temperature simulation signals. In order to improve the accuracy of airflow temperature measurement and reduce the measurement dynamic error. Based on the analysis and research of the airflow temperature signal simulation system, this paper adopts double thermocouples to measure the airflow temperature, and proposes a time domain compensation method to compensate the airflow temperature measurement results. Experimental results show that the double thermocouple compensation method can reduce the measurement error effectively.
{"title":"Dynamic Compensation Method for Temperature Measurement Error of High Temperature and High-speed Airflow","authors":"Jingrui Chu, Shi-kui Wang, Gang Yang, Wenhang Wang","doi":"10.1109/ICMSP55950.2022.9859113","DOIUrl":"https://doi.org/10.1109/ICMSP55950.2022.9859113","url":null,"abstract":"Aiming at the simulation requirements of airflow temperature signal for the total temperature simulation device of hypersonic aircraft, an airflow temperature simulation system is designed. The system consists of a high-temperature plasma spray gun, an airflow mixing chamber, and a thermocouple. The maximal temperature of the system is high and the temperature changes rapidly. But the thermocouple responds slowly, hence there is a large error when measuring the airflow temperature. However, the accuracy of airflow temperature measurement results is important to improve the accuracy of airflow temperature simulation signals. In order to improve the accuracy of airflow temperature measurement and reduce the measurement dynamic error. Based on the analysis and research of the airflow temperature signal simulation system, this paper adopts double thermocouples to measure the airflow temperature, and proposes a time domain compensation method to compensate the airflow temperature measurement results. Experimental results show that the double thermocouple compensation method can reduce the measurement error effectively.","PeriodicalId":114259,"journal":{"name":"2022 4th International Conference on Intelligent Control, Measurement and Signal Processing (ICMSP)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133863342","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 : 2022-07-08DOI: 10.1109/ICMSP55950.2022.9859114
Xudong Feng, Bo Dang, Hongwei Qin
At present, the pipeline transportation of heavy oil mainly adopts heating transportation technology, which consumes a lot of energy and may cause condensation accidents. Therefore, the research and application of pipeline heavy oil cold transportation technology is very necessary. High-power ultrasound can reduce the viscosity of heavy oil to increase its fluidity. This technology has the advantages of low cost and no harm to the environment, and has a good development prospect in the application of heavy oil cold transportation. This paper studies the ultrasonic viscosity reduction technology for heavy oil, and designs the ultrasonic viscosity reduction system based on the analysis of the ultrasonic viscosity reduction mechanism, including the parameters, materials, structure design of the ultrasonic transducer and the circuit design of the ultrasonic generator. The ultrasonic viscosity reduction system was tested on the indoor test platform, and the viscosity reduction effect and influencing factors of the system were analyzed through the test. The experimental results show that the viscosity reduction rate of heavy oil samples can reach more than 40% irreversibly after the system is applied to the heavy oil samples for 20 minutes at room temperature, and the same viscosity reduction effect can be achieved only by using water bath heating method, which requires the water temperature to be above 60°C. Therefore, ultrasonic viscosity reduction system has more advantages than the traditional heating viscosity reduction method, which provides a valuable reference for the application of ultrasonic viscosity reduction technology.
{"title":"Design of viscosity reduction system for heavy oil based on ultrasonic","authors":"Xudong Feng, Bo Dang, Hongwei Qin","doi":"10.1109/ICMSP55950.2022.9859114","DOIUrl":"https://doi.org/10.1109/ICMSP55950.2022.9859114","url":null,"abstract":"At present, the pipeline transportation of heavy oil mainly adopts heating transportation technology, which consumes a lot of energy and may cause condensation accidents. Therefore, the research and application of pipeline heavy oil cold transportation technology is very necessary. High-power ultrasound can reduce the viscosity of heavy oil to increase its fluidity. This technology has the advantages of low cost and no harm to the environment, and has a good development prospect in the application of heavy oil cold transportation. This paper studies the ultrasonic viscosity reduction technology for heavy oil, and designs the ultrasonic viscosity reduction system based on the analysis of the ultrasonic viscosity reduction mechanism, including the parameters, materials, structure design of the ultrasonic transducer and the circuit design of the ultrasonic generator. The ultrasonic viscosity reduction system was tested on the indoor test platform, and the viscosity reduction effect and influencing factors of the system were analyzed through the test. The experimental results show that the viscosity reduction rate of heavy oil samples can reach more than 40% irreversibly after the system is applied to the heavy oil samples for 20 minutes at room temperature, and the same viscosity reduction effect can be achieved only by using water bath heating method, which requires the water temperature to be above 60°C. Therefore, ultrasonic viscosity reduction system has more advantages than the traditional heating viscosity reduction method, which provides a valuable reference for the application of ultrasonic viscosity reduction technology.","PeriodicalId":114259,"journal":{"name":"2022 4th International Conference on Intelligent Control, Measurement and Signal Processing (ICMSP)","volume":"117 12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116590034","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 : 2022-07-08DOI: 10.1109/ICMSP55950.2022.9858947
Ming Zhang, Cuiyun Gao, Siqiang Xu
At present, password recognition in smart home has achieved some research results, but due to the relatively large China's regional differences, dialect password recognition rate in smart home is low. Aiming at the defects of dialect password recognition in smart home, a dialect password recognition system based on Convolutional Neural Network was constructed. In this study, Mel Frequency Cepstrum Coefficient is used to extract the features of Anhui northern dialect, and a Convolutional Neural Network model is constructed for training and recognition. Experimental results show that the model has a high recognition rate for dialect passwords in smart home.
{"title":"Dialect Password Recognition In Smart Home Based On Convolutional Neural Network","authors":"Ming Zhang, Cuiyun Gao, Siqiang Xu","doi":"10.1109/ICMSP55950.2022.9858947","DOIUrl":"https://doi.org/10.1109/ICMSP55950.2022.9858947","url":null,"abstract":"At present, password recognition in smart home has achieved some research results, but due to the relatively large China's regional differences, dialect password recognition rate in smart home is low. Aiming at the defects of dialect password recognition in smart home, a dialect password recognition system based on Convolutional Neural Network was constructed. In this study, Mel Frequency Cepstrum Coefficient is used to extract the features of Anhui northern dialect, and a Convolutional Neural Network model is constructed for training and recognition. Experimental results show that the model has a high recognition rate for dialect passwords in smart home.","PeriodicalId":114259,"journal":{"name":"2022 4th International Conference on Intelligent Control, Measurement and Signal Processing (ICMSP)","volume":"68 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116752648","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 : 2022-07-08DOI: 10.1109/ICMSP55950.2022.9859058
Zhengguo Yan, Yang LI, Jie Liu, Ke Li
With the number of casing damage wells in domestic oil and gas fields increases on a large scale, the problem of low accuracy in casing damage detection of Oil and gas Wells becomes increasingly serious. The Multi Finger Imaging Tool (MIT) and the Magnetic Thickness Tool (MTT) combination logging tool has unique advantages in detecting the change of casing diameter and wall thickness, but it is difficult to reflect the casing damage condition comprehensively and accurately. Considering that the VideoLog downhole TV can combine advantages with the former to make up for shortcomings, and further accurately detect casing loss, this paper proposes a visual combination logging technology scheme of MIT and MTT measurement. MIT/MTT and the VideoLog are mechanically connected by a special conversion head. The communication mode adopts the combination of downhole multi-core bus. By using parallel transmission technology and combining the characteristics of time-sharing operation of the instrument, the problem of signal transmission interference between the two is solved. It provides a set of innovative technical means of low cost, high efficiency and high value for the domestic and overseas field of coat damage detection.
{"title":"MIT/MTT and Visualization Logging Technology and Its Application","authors":"Zhengguo Yan, Yang LI, Jie Liu, Ke Li","doi":"10.1109/ICMSP55950.2022.9859058","DOIUrl":"https://doi.org/10.1109/ICMSP55950.2022.9859058","url":null,"abstract":"With the number of casing damage wells in domestic oil and gas fields increases on a large scale, the problem of low accuracy in casing damage detection of Oil and gas Wells becomes increasingly serious. The Multi Finger Imaging Tool (MIT) and the Magnetic Thickness Tool (MTT) combination logging tool has unique advantages in detecting the change of casing diameter and wall thickness, but it is difficult to reflect the casing damage condition comprehensively and accurately. Considering that the VideoLog downhole TV can combine advantages with the former to make up for shortcomings, and further accurately detect casing loss, this paper proposes a visual combination logging technology scheme of MIT and MTT measurement. MIT/MTT and the VideoLog are mechanically connected by a special conversion head. The communication mode adopts the combination of downhole multi-core bus. By using parallel transmission technology and combining the characteristics of time-sharing operation of the instrument, the problem of signal transmission interference between the two is solved. It provides a set of innovative technical means of low cost, high efficiency and high value for the domestic and overseas field of coat damage detection.","PeriodicalId":114259,"journal":{"name":"2022 4th International Conference on Intelligent Control, Measurement and Signal Processing (ICMSP)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115318605","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 : 2022-07-08DOI: 10.1109/ICMSP55950.2022.9858960
X. Guan, Xinxin Sun
Vehicle identification and traffic statistics at traffic intersections is an effective approach to solve traffic congestion and realize intelligent traffic. This system proposes a vehicle detection and traffic statistics method at complex traffic intersections depending on YOLOV3 and DeepSORT algorithms. The system is suitable for multi-mode traffic junctions and can obtain accurate statistical performance.
{"title":"Vehicle Identification and Traffic Statistics System in Traffic Intersection","authors":"X. Guan, Xinxin Sun","doi":"10.1109/ICMSP55950.2022.9858960","DOIUrl":"https://doi.org/10.1109/ICMSP55950.2022.9858960","url":null,"abstract":"Vehicle identification and traffic statistics at traffic intersections is an effective approach to solve traffic congestion and realize intelligent traffic. This system proposes a vehicle detection and traffic statistics method at complex traffic intersections depending on YOLOV3 and DeepSORT algorithms. The system is suitable for multi-mode traffic junctions and can obtain accurate statistical performance.","PeriodicalId":114259,"journal":{"name":"2022 4th International Conference on Intelligent Control, Measurement and Signal Processing (ICMSP)","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124459602","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 : 2022-07-08DOI: 10.1109/ICMSP55950.2022.9859157
Gaohui Zhou, Songlin Zhang
The key issue of robust M estimation is to construct equivalent weights based on residuals to down-weight outlying observations. However, the correlation of residuals may cause incorrectly down-weighting normal observations. Therefore, this paper aims to propose an M estimation that down-weights outlying observations based on outliers separation. The results of two examples show that the proposed method has high stability. Meanwhile, it can correctly locate the position of outliers even if the outlier rate is as high as 1/3.
{"title":"An M Estimation Based on Outliers Separation","authors":"Gaohui Zhou, Songlin Zhang","doi":"10.1109/ICMSP55950.2022.9859157","DOIUrl":"https://doi.org/10.1109/ICMSP55950.2022.9859157","url":null,"abstract":"The key issue of robust M estimation is to construct equivalent weights based on residuals to down-weight outlying observations. However, the correlation of residuals may cause incorrectly down-weighting normal observations. Therefore, this paper aims to propose an M estimation that down-weights outlying observations based on outliers separation. The results of two examples show that the proposed method has high stability. Meanwhile, it can correctly locate the position of outliers even if the outlier rate is as high as 1/3.","PeriodicalId":114259,"journal":{"name":"2022 4th International Conference on Intelligent Control, Measurement and Signal Processing (ICMSP)","volume":"72 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123604571","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 : 2022-07-08DOI: 10.1109/ICMSP55950.2022.9858942
Chen Kang, Feng Liu
The quality of tension control directly affects the quality of winding products. The quality of tension control directly affects the quality of winding products. At present, most winding systems adopt indirect tension control with dancing roller to meet the requirements in structure, cost and accuracy. Aiming at this goal, this article designs a direct control system based on Siemens 314 PLC and S120 frequency converter by analyzing the winding process and mathematical modeling of tension control. This system adopts a direct closed-loop tension control scheme and converts the tension problem into the control of motor speed. This tension control system has a very friendly man-machine dialogue interface, convenient operation and high reliability, which significantly improves the control accuracy of material winding tension and can be widely used in the actual production of material winding industry.
{"title":"The design of material winding system based on tension control","authors":"Chen Kang, Feng Liu","doi":"10.1109/ICMSP55950.2022.9858942","DOIUrl":"https://doi.org/10.1109/ICMSP55950.2022.9858942","url":null,"abstract":"The quality of tension control directly affects the quality of winding products. The quality of tension control directly affects the quality of winding products. At present, most winding systems adopt indirect tension control with dancing roller to meet the requirements in structure, cost and accuracy. Aiming at this goal, this article designs a direct control system based on Siemens 314 PLC and S120 frequency converter by analyzing the winding process and mathematical modeling of tension control. This system adopts a direct closed-loop tension control scheme and converts the tension problem into the control of motor speed. This tension control system has a very friendly man-machine dialogue interface, convenient operation and high reliability, which significantly improves the control accuracy of material winding tension and can be widely used in the actual production of material winding industry.","PeriodicalId":114259,"journal":{"name":"2022 4th International Conference on Intelligent Control, Measurement and Signal Processing (ICMSP)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121680072","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 : 2022-07-08DOI: 10.1109/ICMSP55950.2022.9858980
Ke Zhang
Aiming at the problems that the existing aerial inspection algorithm in the transmission line is affected by the environment and complex background and other factors, resulting in poor detection effect, low accuracy and poor real-time performance. A recognition algorithm for multi-target detection of transmission lines based on BAM's Faster R-CNN is proposed. Firstly, the structure of the traditional Fast-R-CNN feature extraction network is optimized by using the residual network ResNet50 as its backbone feature extraction network and adding BAM (bottleneck attention module) to the network to improve the visibility and accuracy of the target region in the image. Secondly, we use Softer NMS instead of NMS to improve the non-maximal suppression of the images taken during the inspection. Then combined with the K-means++ clustering algorithm to optimize the anchor parameters. Finally, the FPN feature pyramid network is integrated to further improve the detection accuracy of the algorithm. The experimental results show that the detection speed and accuracy of the improved Faster R-CNN have been effectively improved, achieving a gain of +4.3 mAP.
{"title":"Faster R-CNN Transmission Line Multi-target Detection Based on BAM","authors":"Ke Zhang","doi":"10.1109/ICMSP55950.2022.9858980","DOIUrl":"https://doi.org/10.1109/ICMSP55950.2022.9858980","url":null,"abstract":"Aiming at the problems that the existing aerial inspection algorithm in the transmission line is affected by the environment and complex background and other factors, resulting in poor detection effect, low accuracy and poor real-time performance. A recognition algorithm for multi-target detection of transmission lines based on BAM's Faster R-CNN is proposed. Firstly, the structure of the traditional Fast-R-CNN feature extraction network is optimized by using the residual network ResNet50 as its backbone feature extraction network and adding BAM (bottleneck attention module) to the network to improve the visibility and accuracy of the target region in the image. Secondly, we use Softer NMS instead of NMS to improve the non-maximal suppression of the images taken during the inspection. Then combined with the K-means++ clustering algorithm to optimize the anchor parameters. Finally, the FPN feature pyramid network is integrated to further improve the detection accuracy of the algorithm. The experimental results show that the detection speed and accuracy of the improved Faster R-CNN have been effectively improved, achieving a gain of +4.3 mAP.","PeriodicalId":114259,"journal":{"name":"2022 4th International Conference on Intelligent Control, Measurement and Signal Processing (ICMSP)","volume":"45 10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122566301","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：