Proceedings of the 2019 4th International Conference on Robotics, Control and Automation最新文献

{"title":"Maximum-Linear-Patch based Outlier Detection for Robust Manifold Learning","authors":"C. Du, Hao Chen, Jun Li, N. Jing, Jiangjiang Wu, Songbing Wu","doi":"10.1145/3351180.3351190","DOIUrl":"https://doi.org/10.1145/3351180.3351190","url":null,"abstract":"In this paper, we propose a novel outlier detection method for robust manifold learning. First, we assume that the input high-dimensional data can be represented by an integration of local linear patches, and each patch consists of samples in the local neighborhood that maintains a linear relationship. Then, we use a hierarchical divisive clustering method to seek maximum linear patches (MLPs) and present a local density based scheme to detect outliers in each MLP. In order to evaluate the performance of outlier detection, we also propose an improved outlier detection evaluation method based on manifold distance, which is suitable for robust manifold learning. Last, we give several experiments to demonstrate the effectiveness of the proposed outlier detection method and evaluation method.","PeriodicalId":375806,"journal":{"name":"Proceedings of the 2019 4th International Conference on Robotics, Control and Automation","volume":"80 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126381348","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}

{"title":"Binocular Stereo Matching Based on Convolutional Neural Networks","authors":"Shuigen Lu, Hesheng Yin, Yunliang Zhu, X. Yang, Shaomiao Li, Bo Huang","doi":"10.1145/3351180.3351189","DOIUrl":"https://doi.org/10.1145/3351180.3351189","url":null,"abstract":"For the binocular stereo matching of deep learning based on patches, the networks structure is vital for matching cost in stereo matching. The task of using a pair of stereo images to estimate depth information can be achieved by a convolutional neural network after being formatted as a supervised learning task. However, the current stereo matching neural networks have poor stereo matching results in ill-posed-regions. In order to solve this problem, Our proposed a deep learning architecture that constructs a cost volume through improving the relationship between groups. The network consists of a feature extraction module, a cross-form spatial pyramid module and a feature matching fusion module. The improved stereo matching network is trained and verified on the KITTI data. The experimental results show that the improved network has certain advantages in terms of accuracy and speed compared with the previous methods.","PeriodicalId":375806,"journal":{"name":"Proceedings of the 2019 4th International Conference on Robotics, Control and Automation","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124248225","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}

{"title":"RGBD Semantic Segmentation Based on Global Convolutional Network","authors":"Xiaoning Gao, Jun Yu, Jian-xun Li","doi":"10.1145/3351180.3351182","DOIUrl":"https://doi.org/10.1145/3351180.3351182","url":null,"abstract":"Convolutional neural networks have gradually dominated the field of image semantic segmentation, and have achieved good results in 2D image semantic segmentation tasks. However, the 2D semantic segmentation algorithm based on CNN is still unsatisfactory in many complex scenarios, such as indoor scenes. Fortunately, advances in depth sensor technology have made it easy to obtain depth information, which carries rich geometric structure information. In order to effectively embed the depth map into the convolutional neural network, this paper introduces the dual encoder fusion network framework to fully exploit the geometric features. For the problem of weakening the local pixel classification ability of the dual encoder fusion network, this paper introduces global convolutional network (GCN), which is based on the large kernel idea, to improve the performance of dual encoder fusion network. Extensive experiments on the NYU v2 dataset demonstrate that the two-encoder fusion network based on global convolution network has much better precision than the original fusion network, and the classification ability for local pixels is stronger.","PeriodicalId":375806,"journal":{"name":"Proceedings of the 2019 4th International Conference on Robotics, Control and Automation","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122612629","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}

{"title":"Research on Three-dimensional Measurement Method Based on Monocular Vision","authors":"Shuai Xue, Xisheng Li, Qing Liu","doi":"10.1145/3351180.3351223","DOIUrl":"https://doi.org/10.1145/3351180.3351223","url":null,"abstract":"3D measurement and reconstruction techniques are widely used in many fields, such as automated production, computer vision, and medical diagnostics. Fast and accurate three-dimensional reconstruction of the surface of the object for three-dimensional reconstruction is an important technical problem [1]. Based on the monocular vision measurement technology, this paper proposes a single-camera single-light source system for 3D reconstruction. The experiment uses a hemisphere of known radius to make the center of the hemisphere coincide with the target center. The circular target center is derived by the algorithm proposed in this paper and three-dimensional information of the location and error analysis.","PeriodicalId":375806,"journal":{"name":"Proceedings of the 2019 4th International Conference on Robotics, Control and Automation","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126561749","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}

{"title":"An Improved NSCT of Fusion of Visible and Infrared Images","authors":"Zeng Xiangjin, Chen Jian, L. Ke","doi":"10.1145/3351180.3351209","DOIUrl":"https://doi.org/10.1145/3351180.3351209","url":null,"abstract":"An improved NSCT of infrared and visible light image fusion method is proposed, which aims at the problem of low contrast between target and background in infrared and visible light fusion images. Firstly, the infrared and visible images are preprocessed respectively. Secondly, the morphological Top-hat operator is used to suppress the background of the image. According to the contrast between the target and the surrounding background, the salient region of interest is segmented. Finally, the high frequency and low frequency components of the visible and infrared images are obtained by NSCT transform. Then, different fusion strategies are designed in the salient and non-salient regions to fuse the low-frequency coefficients and high-frequency coefficients. Experimental results show that compared with other algorithms, the fusion image obtained by this method enhances the contrast between the thermal target and the background, while retaining the details of the background information, more in line with human visual perception.","PeriodicalId":375806,"journal":{"name":"Proceedings of the 2019 4th International Conference on Robotics, Control and Automation","volume":"51 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127007112","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}

{"title":"Application of Improved VIBE Algorithm in Robot Grabbing System Based on Visual Servo","authors":"Shuangshuang Liu, Jin Huang, Yaogao Shen, Zichang Guo","doi":"10.1145/3351180.3351200","DOIUrl":"https://doi.org/10.1145/3351180.3351200","url":null,"abstract":"Object detection is the essential part for arm robot grabbing system. Traditional VIBE (Visual Background Extractor) algorithm is used to capture moving objects in a visual servo-based robot grabbing system, but there are three major problems in the image: ghost, shadow and hole. Aiming at the three defects of VIBE algorithm, the corresponding improvement schemes are put forward: the frame difference method combined with VIBE algorithm is designed to solve the 'ghost' problem, the improved model based on normalized RGB space is used to eliminate the 'shadow'. Finally, image enclosure algorithm is used to deal with \"holes\". Through the comparative analysis of the experimental results, the effectiveness of the improved algorithm is verified, and the above three problems are well solved.","PeriodicalId":375806,"journal":{"name":"Proceedings of the 2019 4th International Conference on Robotics, Control and Automation","volume":"151 3","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114044755","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}

{"title":"Distributed Robust Filtering in Sensor Network with Random Communication Delays","authors":"Fengzeng Zhu, Li Peng","doi":"10.1145/3351180.3351214","DOIUrl":"https://doi.org/10.1145/3351180.3351214","url":null,"abstract":"In this paper, the problem of l2-l∞ filters design is addressed for sensor networks with one-step random communication delay. In a distributed filter network, each local filter that estimates the state of the system not only utilizes the information obtained by the node itself, but also utilizes the information of the neighbor nodes. The random communication delays of measurement are described by a binary switching sequence satisfying a conditional probability distribution. Subsequently, by constructing the Lyapunov function, sufficient conditions are obtained to ensure that the filter error dynamic system has exponential mean-square stability and l2-l∞ performance. Then, the parameters of the distributed l2-l∞ filter are obtained by a linear matrix inequality (LMI) technique. Finally, the effectiveness of the method is proved by a numerical simulation example.","PeriodicalId":375806,"journal":{"name":"Proceedings of the 2019 4th International Conference on Robotics, Control and Automation","volume":"546 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116459246","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}

{"title":"Research on Formation Algorithm Based on Second-order Delay Multi-Agent System","authors":"Huarong Zhao, Li Peng, Fengzeng Zhu","doi":"10.1145/3351180.3351224","DOIUrl":"https://doi.org/10.1145/3351180.3351224","url":null,"abstract":"This paper considers the time delay issue in the communication process among multi-agents with directed fixed topology and presents a consensus protocol. Then a sufficient condition for multi-agent system obtaining global asymptotic consistency under directed fixed topological is proved. Furthermore, the upper bound of compactness for the largest time delay is provided and the information consistency idea is applied to the formation control of multi-agent. Finally, some simulations are performed to verify the theoretical results proposed in this paper and the simulation results show that method based on the consistency protocol can be successfully applied to multi-agent networked system with communication delay and the preset formation can be completed well.","PeriodicalId":375806,"journal":{"name":"Proceedings of the 2019 4th International Conference on Robotics, Control and Automation","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123213939","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}

{"title":"Pre-detection and Type Discrimination of Human Trembling Diseases","authors":"Fan-Sin Chen, Jie Li, Deng-You Lyu, Tsorng-Lin Chia, C. Chia, Yi Wu","doi":"10.1145/3351180.3351188","DOIUrl":"https://doi.org/10.1145/3351180.3351188","url":null,"abstract":"The slight tremor of the human body is an early sign and symptom of many diseases. If it can be found at the beginning of the tremor, it will be beneficial for subsequent medical treatment. In this paper, we use deep learning-based motion magnification to analyze subtle changes in the image that cannot be observed by the human eye and to filter and amplify the frequencies of different trembling video for achieving prediction for different types of tremors. It can effectively reduce the medical evaluation time of tremble detection. The results of a video containing human respiration and tremor indicate that the proposed method can discriminate different tremors.","PeriodicalId":375806,"journal":{"name":"Proceedings of the 2019 4th International Conference on Robotics, Control and Automation","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123534503","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}

{"title":"An Intelligent Identification and Acquisition System for UAVs Based on Edge Computing Using in the Transmission Line Inspection","authors":"Liu Yue, Wang Wanguo, Xu Ronghao, Li Zengwei, Tian Ran Yuan","doi":"10.1145/3351180.3351186","DOIUrl":"https://doi.org/10.1145/3351180.3351186","url":null,"abstract":"With the increase of the transmission line mileage, the transmission line inspection is facing severe challenges. As an efficient inspection way, the UAV (Unnamed Aerial Vechicle) is used for the transmission line inspection. However, at the present stage, to complete the observation and data collection of transmission line equipment, the PTZ (Pan Tilt Zoom) camera in the UAV is completely controlled dependent on the operator during the inspection. These needs higher requirements for the operator and lead to high labor intensity. And the image scale of the collection device is not uniform, which improves the difficulty of subsequent intelligent diagnosis of defects. In order to solve this problem, this paper designs an intelligent acquisition system for UAVs based on edge computing using in the transmission line inspection. The system includes the front-end edge computing detection module based on SSD (Single Shot Multibox Detector) algorithm and a PTZ camera control module. The system workflow is as follows: firstly, the UAV acquires the video stream data of the surrounding environment of the transmission line through the fixed-focus camera in the PTZ. The front-end edge computing detection module receives the video stream through the USB interface. And the front-end edge computing module uses the SSD algorithm to identify the video stream gotten from the PTZ camera to recognize and locate the transmission line equipment. Secondly, if the transmission line equipment is identified in the field of view, such as an insulator, the anti-vibration hammer, the equalizing ring, etc., the front-end edge computing module calculates the target image ratio of the device and from the center position based on the identified pixel position information of the transmission line equipment, and converts to the pan-tilt rotation amount and camera pulling factor. The pan-tilt rotation amount and camera pulling factor is transmitted to the PTZ camera control module through the serial port. The PTZ camera control module controls the camera's movement until the transmission line equipment is in the center of the field of view. Finally, the wide-angle camera in dual-view PTZ camera is controlled to zoom. And the image after the zoom is photographed and saved to finish a flow of the data of transmission line equipment collects. After testing, the SSD algorithm for transmission line equipment can achieve an overall recognition accuracy of 73%. The insulator identification accuracy is 75.7%.The anti-vibration hammer identification accuracy is 72.6%.And the equalization ring recognition accuracy is 70.3%.The acquisition video is 1080P, the edge computing module processing speed is 3fp/s, and the pan/tilt control module controls the pan/tilt camera to complete a shooting time of 3s. On the one hand, the system realizes the autonomy, process and standardized collection of the image information during the transmission line equipment while improving the validity of the image da","PeriodicalId":375806,"journal":{"name":"Proceedings of the 2019 4th International Conference on Robotics, Control and Automation","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122579279","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}