However, the scope and distance of optical imaging were limited, especially in the case of muddy water, the propagation of optical information was seriously interfered, and imaging became more difficult. Due to the complex and changeable underwater environment and the nature of acoustic imaging, sonar image has noise, low resolution and fuzzy details, which has a great impact on the recognition and interpretation of sonar image. On the basis of the original SRGAN network, this paper improves and optimates its network structure and loss function. Replace the ordinary convolution layer with the void convolution layer in the residual block structure of the generated network, delete the batch normalization layer (BN layer), reduce the resource consumption and expand the receiver field, so as to improve the training efficiency of the network; A gradient penalty term is added to the improved discriminant network loss function to accelerate the convergence of the network and improve the stability of training. Four classical image super resolution algorithms are compared with the improved SRGAN algorithm under the verification of sonar dataset. The experimental results show that the improved SRGAN network is superior to the traditional super resolution method in the reconstruction of sonar image in terms of rich texture and details, and improves the quality of sonar image super resolution reconstruction.
{"title":"A super-resolution reconstruction method of underwater target detection image by side scan sonar","authors":"Jin Hua, Mengzhao Liu, Shujia Wang","doi":"10.1145/3483845.3483869","DOIUrl":"https://doi.org/10.1145/3483845.3483869","url":null,"abstract":"However, the scope and distance of optical imaging were limited, especially in the case of muddy water, the propagation of optical information was seriously interfered, and imaging became more difficult. Due to the complex and changeable underwater environment and the nature of acoustic imaging, sonar image has noise, low resolution and fuzzy details, which has a great impact on the recognition and interpretation of sonar image. On the basis of the original SRGAN network, this paper improves and optimates its network structure and loss function. Replace the ordinary convolution layer with the void convolution layer in the residual block structure of the generated network, delete the batch normalization layer (BN layer), reduce the resource consumption and expand the receiver field, so as to improve the training efficiency of the network; A gradient penalty term is added to the improved discriminant network loss function to accelerate the convergence of the network and improve the stability of training. Four classical image super resolution algorithms are compared with the improved SRGAN algorithm under the verification of sonar dataset. The experimental results show that the improved SRGAN network is superior to the traditional super resolution method in the reconstruction of sonar image in terms of rich texture and details, and improves the quality of sonar image super resolution reconstruction.","PeriodicalId":134636,"journal":{"name":"Proceedings of the 2021 2nd International Conference on Control, Robotics and Intelligent System","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123360193","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 recent years, peer-to-peer (P2P) system has been widely used in the Internet. The security research for the network will provide guarantee for its stable operation. Based on Gephi platform, this paper deeply studies the complex P2P network topology, and gives the countermeasures and suggestions of network protection. The specific method is: take the complex P2P network of Gnutella protocol as the sample, through the Gephi platform, this paper analyzes the network topology node’s in degree, out degree and overall degree, and gives the network node’s protection strategy. The results show that although the nodes in P2P network do not distinguish between server and client, each node still tends to play one of the roles, and the network shows an obvious community structure, so the network can still adopt focused protection strategies.
{"title":"Security Analysis of Peer-to-Peer Network Topology Based on Gephi Platform: P2P Network Security Analysis via Gephi","authors":"Xue Zheng, Jun-zhong He, Xiaojing Gao, Kebin Chen","doi":"10.1145/3483845.3483876","DOIUrl":"https://doi.org/10.1145/3483845.3483876","url":null,"abstract":"∗In recent years, peer-to-peer (P2P) system has been widely used in the Internet. The security research for the network will provide guarantee for its stable operation. Based on Gephi platform, this paper deeply studies the complex P2P network topology, and gives the countermeasures and suggestions of network protection. The specific method is: take the complex P2P network of Gnutella protocol as the sample, through the Gephi platform, this paper analyzes the network topology node’s in degree, out degree and overall degree, and gives the network node’s protection strategy. The results show that although the nodes in P2P network do not distinguish between server and client, each node still tends to play one of the roles, and the network shows an obvious community structure, so the network can still adopt focused protection strategies.","PeriodicalId":134636,"journal":{"name":"Proceedings of the 2021 2nd International Conference on Control, Robotics and Intelligent System","volume":"84 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133109207","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}
For some large and complex operating platforms in machinery factories, in order to detect fatigued workers in time, machine vision is adopted for fatigue detection. Normal camera cannot cover the entire working range. So come up with an idea to use an industrial computer to control the pan-tilt for face tracking and fatigue detection. First, face detection is performed on the image information captured by the pan-tilt. Then, the pan-tilt is controlled to rotate according to the position of the face to achieve the purpose of tracking the face of the staff in real time. On this basis, the feature points recognized by face detection algorithm and the PERCLOS algorithm are used to calculate, The thresholds of EAR value and MAR value were set as 0.18 and 0.4 respectively according to the experimental results to identify the fatigue features of the face for fatigue detection. Finally, whether the person is in a state of fatigue is judged according to whether the percentage of the time that the person is in a fatigue characteristic exceeds 75% .The results show that the corresponding hardware equipment and the algorithm used can make the recognition accuracy basically reach 90% during working time, and the detection time is less than 90ms, which satisfy the requirements of real-time face tracking and fatigue detection accuracy and execution efficiency under working state.
{"title":"Detection of Fatigued Face","authors":"Ming-da Song, Song Xin, Ze-ming Wang, Gang Zhao","doi":"10.1145/3483845.3483885","DOIUrl":"https://doi.org/10.1145/3483845.3483885","url":null,"abstract":"For some large and complex operating platforms in machinery factories, in order to detect fatigued workers in time, machine vision is adopted for fatigue detection. Normal camera cannot cover the entire working range. So come up with an idea to use an industrial computer to control the pan-tilt for face tracking and fatigue detection. First, face detection is performed on the image information captured by the pan-tilt. Then, the pan-tilt is controlled to rotate according to the position of the face to achieve the purpose of tracking the face of the staff in real time. On this basis, the feature points recognized by face detection algorithm and the PERCLOS algorithm are used to calculate, The thresholds of EAR value and MAR value were set as 0.18 and 0.4 respectively according to the experimental results to identify the fatigue features of the face for fatigue detection. Finally, whether the person is in a state of fatigue is judged according to whether the percentage of the time that the person is in a fatigue characteristic exceeds 75% .The results show that the corresponding hardware equipment and the algorithm used can make the recognition accuracy basically reach 90% during working time, and the detection time is less than 90ms, which satisfy the requirements of real-time face tracking and fatigue detection accuracy and execution efficiency under working state.","PeriodicalId":134636,"journal":{"name":"Proceedings of the 2021 2nd International Conference on Control, Robotics and Intelligent System","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114881580","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}
Carried out in-depth research for the parallel mechanism 2P3RR, introduced composition of the mechanism, derived the kinematics formulas which include the positive and the inverse solutions of the position, and the Jacobian matrix of the mechanism is obtained. The parallel mechanism's characteristics such as workspace and singularity are analyzed. The workspace of the mechanism is analyzed by graphic method, and the singularity of Jacobian matrix is used to analyze three singular forms of the mechanism. It can be seen that the positive and inverse kinematics solutions have explicit expressions by solving the parallel mechanism kinematics, the workspace is easy to describe and the singularity of the mechanism can be avoided, which lays a theoretical foundation for the design and development of the mechanism and its application in the field of robots and machine tools.
{"title":"Workspace and Singularity Analysis of a 2P3RR Planar Parallel Mechanism","authors":"Xuzhao Han, Peng Jin, Suolao Li, Wen Liu, Yuanxiong Zhang, Yi Guo","doi":"10.1145/3483845.3483851","DOIUrl":"https://doi.org/10.1145/3483845.3483851","url":null,"abstract":"Carried out in-depth research for the parallel mechanism 2P3RR, introduced composition of the mechanism, derived the kinematics formulas which include the positive and the inverse solutions of the position, and the Jacobian matrix of the mechanism is obtained. The parallel mechanism's characteristics such as workspace and singularity are analyzed. The workspace of the mechanism is analyzed by graphic method, and the singularity of Jacobian matrix is used to analyze three singular forms of the mechanism. It can be seen that the positive and inverse kinematics solutions have explicit expressions by solving the parallel mechanism kinematics, the workspace is easy to describe and the singularity of the mechanism can be avoided, which lays a theoretical foundation for the design and development of the mechanism and its application in the field of robots and machine tools.","PeriodicalId":134636,"journal":{"name":"Proceedings of the 2021 2nd International Conference on Control, Robotics and Intelligent System","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123958471","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}
Jing Zhou, Xiaozhe Zhao, Zhen Xu, Si-jun Peng, Zhong Lin
Task allocation is a key part of unmanned aerial vehicle (UAV) swarm. Although a large number of solving algorithms have been developed, there are few technologies that support task allocation algorithms in dynamic environments. Obviously, this is not in accord with the actual situation. The battlefield is changing rapidly, which may lead to the failure of the allocated tasks and the inability to allocate new tasks. In order to deal with this problem, this paper improves the original D-NSGA3 algorithm to adapt to the dynamic environment. The experimental results show that, compared with the original static algorithm, the proposed algorithm has better effect in solving the task allocation problem of high-dimensional multi-objective agent based on maximizing the number of successfully allocated tasks, maximizing the benefits of executing tasks, minimizing the consumption cost, and minimizing the time cost.
{"title":"Dynamic task allocation algorithm based on D-NSGA3","authors":"Jing Zhou, Xiaozhe Zhao, Zhen Xu, Si-jun Peng, Zhong Lin","doi":"10.1145/3483845.3483875","DOIUrl":"https://doi.org/10.1145/3483845.3483875","url":null,"abstract":"Task allocation is a key part of unmanned aerial vehicle (UAV) swarm. Although a large number of solving algorithms have been developed, there are few technologies that support task allocation algorithms in dynamic environments. Obviously, this is not in accord with the actual situation. The battlefield is changing rapidly, which may lead to the failure of the allocated tasks and the inability to allocate new tasks. In order to deal with this problem, this paper improves the original D-NSGA3 algorithm to adapt to the dynamic environment. The experimental results show that, compared with the original static algorithm, the proposed algorithm has better effect in solving the task allocation problem of high-dimensional multi-objective agent based on maximizing the number of successfully allocated tasks, maximizing the benefits of executing tasks, minimizing the consumption cost, and minimizing the time cost.","PeriodicalId":134636,"journal":{"name":"Proceedings of the 2021 2nd International Conference on Control, Robotics and Intelligent System","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124570485","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}
Aiming at the problem that traditional forward collision warning (FCW) system cannot deal with some emergencies (such as the sudden appearance of target) accurately and timely in the real world, we present MIFWS-FCW- a novel system that applies multi-information fusion and improved warning strategy to improve the accuracy of emergencies warning. In this work, we construct two dynamic safety warning areas changing with the speed of the vehicle according to the detected lane lines and the safety distance model. We predict the trajectory of target based on the information output by YOLOv3 (You Only Look Once v3) at the same time. Then, three-level warnings are designed based on the improved warning strategy, and we combine the distance perception to decide whether the system sends out strong reminders. We transplant our system to Atlas 200 Developer Kit and conduct quantitative evaluation on 20 videos with different environment. The experimental results show that the MIFWS-FCW system can achieve the competitive warning accuracy of 95.5% with the warning time of 1.16s, to some extent, it can reduce the incidence of accidents and achieve the purpose of auxiliary driving.
针对传统的前向碰撞预警系统(FCW)不能准确、及时地处理现实世界中某些突发事件(如目标突然出现)的问题，提出了一种应用多信息融合和改进预警策略的MIFWS-FCW系统，以提高突发事件预警的准确性。在这项工作中，我们根据检测到的车道线和安全距离模型构建了两个随车辆速度变化的动态安全预警区域。我们根据YOLOv3 (You Only Look Once v3)同时输出的信息来预测目标的轨迹。然后，基于改进的预警策略设计三级预警，并结合距离感知来决定系统是否发出强预警。我们将我们的系统移植到Atlas 200 Developer Kit上，对20个不同环境下的视频进行了定量评估。实验结果表明，MIFWS-FCW系统在1.16s的预警时间内，竞争预警准确率达到95.5%，在一定程度上降低了事故发生率，达到了辅助驾驶的目的。
{"title":"Vehicle Forward Collision Warning Algorithm Based on Multi-Information Fusion and Improved Warning Strategy","authors":"Yining Pan, Yanliang Jin, Rukun Lyu","doi":"10.1145/3483845.3483892","DOIUrl":"https://doi.org/10.1145/3483845.3483892","url":null,"abstract":"Aiming at the problem that traditional forward collision warning (FCW) system cannot deal with some emergencies (such as the sudden appearance of target) accurately and timely in the real world, we present MIFWS-FCW- a novel system that applies multi-information fusion and improved warning strategy to improve the accuracy of emergencies warning. In this work, we construct two dynamic safety warning areas changing with the speed of the vehicle according to the detected lane lines and the safety distance model. We predict the trajectory of target based on the information output by YOLOv3 (You Only Look Once v3) at the same time. Then, three-level warnings are designed based on the improved warning strategy, and we combine the distance perception to decide whether the system sends out strong reminders. We transplant our system to Atlas 200 Developer Kit and conduct quantitative evaluation on 20 videos with different environment. The experimental results show that the MIFWS-FCW system can achieve the competitive warning accuracy of 95.5% with the warning time of 1.16s, to some extent, it can reduce the incidence of accidents and achieve the purpose of auxiliary driving.","PeriodicalId":134636,"journal":{"name":"Proceedings of the 2021 2nd International Conference on Control, Robotics and Intelligent System","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130573315","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}
∗Because of its lightweight, small size, short launch cycle, and many other advantages, microsatellite has been widely used in military and civil fields. In this paper, the attitude control system of a microsatellite is designed according to the project’s requirements, which provides a reference for others to research microsatellite attitude control. This paper first introduces the system’s hardware composition and working mode according to the criteria of attitude control, then designs the software operation flow, data acquisition, and control sequence of the attitude control system, and finally carries out the simulation test to verify the rationality of the design.
{"title":"Design and Verification of Attitude Control System for A Small Satellite","authors":"Shuyan Ni, Shimiao Chen, Yurong Liao, Naiping Cheng","doi":"10.1145/3483845.3483861","DOIUrl":"https://doi.org/10.1145/3483845.3483861","url":null,"abstract":"∗Because of its lightweight, small size, short launch cycle, and many other advantages, microsatellite has been widely used in military and civil fields. In this paper, the attitude control system of a microsatellite is designed according to the project’s requirements, which provides a reference for others to research microsatellite attitude control. This paper first introduces the system’s hardware composition and working mode according to the criteria of attitude control, then designs the software operation flow, data acquisition, and control sequence of the attitude control system, and finally carries out the simulation test to verify the rationality of the design.","PeriodicalId":134636,"journal":{"name":"Proceedings of the 2021 2nd International Conference on Control, Robotics and Intelligent System","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125459589","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 selects the outer contour method to describe the image's features and uses industrial robots to draw. Based on the minimum description length, the Gaussian filter in the Canny operator is improved adaptively to ensure edge recognition accuracy. To highlight the primary contour in the process of drawing the image, delete the small areas. We propose an edge thinning algorithm to avoid drawing the same contour twice and obtain the edges with single-pixel width. The broken contours are connected to reduce the time of lifting the pen when the robot is drawing. To avoid contour tracking error resulting in incomplete image rendering, removing the burr points in the image. Finally, we use Robotstudio simulation software and ABB IRB1200 robot to simulate and experiment with the obtained contour image. The results show that the proposed method can effectively filter out noise and irrelevant details. We realize the goal of edge thinning and improve the integrity of each section contour. Reduce the number of robot lift pens, and improve the efficiency of drawing.
{"title":"Research on Drawing Robot Based on Image Edge Detection","authors":"Feng Liu, L. Cao, Zibo Sun, Zheng Li","doi":"10.1145/3483845.3483847","DOIUrl":"https://doi.org/10.1145/3483845.3483847","url":null,"abstract":"This paper selects the outer contour method to describe the image's features and uses industrial robots to draw. Based on the minimum description length, the Gaussian filter in the Canny operator is improved adaptively to ensure edge recognition accuracy. To highlight the primary contour in the process of drawing the image, delete the small areas. We propose an edge thinning algorithm to avoid drawing the same contour twice and obtain the edges with single-pixel width. The broken contours are connected to reduce the time of lifting the pen when the robot is drawing. To avoid contour tracking error resulting in incomplete image rendering, removing the burr points in the image. Finally, we use Robotstudio simulation software and ABB IRB1200 robot to simulate and experiment with the obtained contour image. The results show that the proposed method can effectively filter out noise and irrelevant details. We realize the goal of edge thinning and improve the integrity of each section contour. Reduce the number of robot lift pens, and improve the efficiency of drawing.","PeriodicalId":134636,"journal":{"name":"Proceedings of the 2021 2nd International Conference on Control, Robotics and Intelligent System","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130628372","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}
Ying Zhao, Yang He, Jiaqi Jing, Tie Wang, Guangmiao Jiang
In recent years, there have been many achievements in Mongolian studies. Scholars have carried out eye movement characteristics analysis and research from the aspects of reading gaze position effect, combined with the effect of high and low frequency words on gaze position and other aspects of Mongolian cognitive research. However, there is no systematic study on the influence of visual angle on Mongolian character recognition. This research addresses the above-mentioned problems, taking visual angle as the independent variable and reaction time as the dependent variable, selecting the commonly used Mongolian vowel consonants and divide them into seven visual angles of 2°, 4°, 7°, 10°, 13°, 16°, and 19°. Discussing the impact of the reaction time on the visual angle and the best recognition visual angle. The research showed that at 19°, the reaction time was the shortest and the recognition effect was better. Therefore, the best visual angle of Mongolian language could be identified in order to improve the efficiency of recognition, enrich the research in the field of Mongolian language study, and promote the development of Mongolian language research.
{"title":"The influence of visual angle on Mongolian character recognition","authors":"Ying Zhao, Yang He, Jiaqi Jing, Tie Wang, Guangmiao Jiang","doi":"10.1145/3483845.3483877","DOIUrl":"https://doi.org/10.1145/3483845.3483877","url":null,"abstract":"In recent years, there have been many achievements in Mongolian studies. Scholars have carried out eye movement characteristics analysis and research from the aspects of reading gaze position effect, combined with the effect of high and low frequency words on gaze position and other aspects of Mongolian cognitive research. However, there is no systematic study on the influence of visual angle on Mongolian character recognition. This research addresses the above-mentioned problems, taking visual angle as the independent variable and reaction time as the dependent variable, selecting the commonly used Mongolian vowel consonants and divide them into seven visual angles of 2°, 4°, 7°, 10°, 13°, 16°, and 19°. Discussing the impact of the reaction time on the visual angle and the best recognition visual angle. The research showed that at 19°, the reaction time was the shortest and the recognition effect was better. Therefore, the best visual angle of Mongolian language could be identified in order to improve the efficiency of recognition, enrich the research in the field of Mongolian language study, and promote the development of Mongolian language research.","PeriodicalId":134636,"journal":{"name":"Proceedings of the 2021 2nd International Conference on Control, Robotics and Intelligent System","volume":"97 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132367035","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 order to reduce the impact of the increasingly frequent sandstorm weather on computer vision-related applications, we propose an enhancement algorithm for sandstorm images. First, we achieve white balance through the Shades of Gray (SoG) algorithm optimized by blue channel compensation. Then, we convert the corrected image into Lab color space. Next, using Contrast Limited Adaptive Histogram Equalization (CLAHE) to enhance the brightness part (L component) to improve the contrast of the sandstorm images. Finally, the L component is further Laplace sharpened to enhance the details. The experimental results show that the proposed method can effectively restore the fade characteristics of sandstorm-degraded images and improve the clarity of the images. In addition, our experiments also prove that our algorithm can significantly improve the feature point extraction results of images captured during sandstorm weather conditions to ameliorate computer vision applications during sandstorm weather.
{"title":"Enhancement Algorithm for Sandstorm Images","authors":"Zhong Hua, B. Nie","doi":"10.1145/3483845.3483893","DOIUrl":"https://doi.org/10.1145/3483845.3483893","url":null,"abstract":"In order to reduce the impact of the increasingly frequent sandstorm weather on computer vision-related applications, we propose an enhancement algorithm for sandstorm images. First, we achieve white balance through the Shades of Gray (SoG) algorithm optimized by blue channel compensation. Then, we convert the corrected image into Lab color space. Next, using Contrast Limited Adaptive Histogram Equalization (CLAHE) to enhance the brightness part (L component) to improve the contrast of the sandstorm images. Finally, the L component is further Laplace sharpened to enhance the details. The experimental results show that the proposed method can effectively restore the fade characteristics of sandstorm-degraded images and improve the clarity of the images. In addition, our experiments also prove that our algorithm can significantly improve the feature point extraction results of images captured during sandstorm weather conditions to ameliorate computer vision applications during sandstorm weather.","PeriodicalId":134636,"journal":{"name":"Proceedings of the 2021 2nd International Conference on Control, Robotics and Intelligent System","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128093709","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：