Partial point cloud registration is an essential preprocessing technique to generate complete 3D shapes that aim to transform partial scans into a common coordinate system. Existing methods that utilize geometric alig...
Partial point cloud registration is an essential preprocessing technique to generate complete 3D shapes that aim to transform partial scans into a common coordinate system. Existing methods that utilize geometric alignment rarely combine local and global information, which results in many incorrect correspondences. In the paper, a two-layer descriptor and a high-dimension searching approach (DDHSM) are utilized to extract and select the correct correspondences. Firstly, in the first layer, the Fast Point Feature Histogram (FPFH) is used to describe the feature of every point with high-dimension information based on its robustness and fastness. Then a high-dimensional search is considered to extract the initial corresponding set. In the second layer, global information is introduced to find the correct correspondences. For each point pair in the initial corresponding set, the normal angles are compared for all remaining points that satisfy the distance requirement. Finally, the singular value decomposition (SVD) method is implemented to compute the rigid transformation based on the updated corresponding set. Experimental results and comparisons with state-of-the-art methods demonstrate the effectiveness and feasibility of our approach.
Multi-Agent Path Finding is a problem of finding the optimal set of paths for multiple agents from the starting position to the goal without conflict, which is essential to large-scale robotic systems. Imitation and r...
详细信息
The deep convolution method based on MSDP signal imaging has been proven to be an effective means of monitoring the robot grinding process. This method has very high requirements on the quality of imaging and requires...
详细信息
ISBN:
(数字)9798350368604
ISBN:
(纸本)9798350368611
The deep convolution method based on MSDP signal imaging has been proven to be an effective means of monitoring the robot grinding process. This method has very high requirements on the quality of imaging and requires a lot of computing resources. This paper proposes an image self-contrastive learning network that can effectively extract fault information from low-pixel fuzzy images and achieve efficient and accurate condition monitoring. The network first calculates the correlation distance matrix between two identical modified symmetrized dot pattern (MSDP) images and uses the generated similarity labels to calculate the contrast loss to extract the embedded features of the image. Finally, the classification network is trained using the cross entropy loss to achieve the classification of fault images. The method was verified on a robotic grinding platform to significantly reduce the calculation time while maintaining the recognition accuracy, which provides reliable technical support for the real-time monitoring of the robotic grinding process.
For quadrotors, imposing multiple dynamic constraints on the state simultaneously to achieve safe control is a challenging problem. In this paper, a cascaded control archi-tecture based on quadratic programming method...
详细信息
The perception capability of robotic systems relies on the richness of the dataset. Although Segment Anything Model 2 (SAM2), trained on large datasets, demonstrates strong perception potential in perception tasks, it...
详细信息
Aiming at the charging and navigation strategy of electric vehicles in the road-electricity coupling scenario, this paper proposes a hybrid planning travel scheme based on the pre-charging and charging warning model b...
Aiming at the charging and navigation strategy of electric vehicles in the road-electricity coupling scenario, this paper proposes a hybrid planning travel scheme based on the pre-charging and charging warning model based on the road traffic congestion predicted by the big data of the Internet of Vehicles. This model integrally balances the interests of users and power grid in the coupled road-electric network, and solves the optimal path scheme under different travel scenarios based on collaborative filtering algorithm and user feedback, considering the influence of vehicle departure time and remaining power in short time scale. It is experimentally demonstrated that the proposed route hybrid recommendation model has better planning effect compared with the single travel model.
As an crucial component of power line, the timely check for the state of insulator is necessary for the normal running of transmission lines. In view of debasement of insulator segmentation accuracy in current transmi...
详细信息
ISBN:
(纸本)9781665479691
As an crucial component of power line, the timely check for the state of insulator is necessary for the normal running of transmission lines. In view of debasement of insulator segmentation accuracy in current transmission line caused by complex background, low contrast, and the quality of images is not guaranteed, we improve U-Net by combining attention mechanism and residual connection. Residual connection is added to the encoder part to improve the extraction of low-level semantic information, and the attention mechanism is added to the decoder part for integrating high and low level features better and reduce the error between them. We confirm the effectiveness of the improved module through experiments, while the results showing that the improved U-Net model segmentation performance on insulator dataset is improved from 0.875 to 0.912. And our method also outperforms previous segmentation work on insulator dataset.
The working state of fasteners is closely related to the safe and stable operation of the power grid. However, the performance of the existing fastener defect detection models is not enough to meet the requirements, b...
详细信息
ISBN:
(纸本)9781665479691
The working state of fasteners is closely related to the safe and stable operation of the power grid. However, the performance of the existing fastener defect detection models is not enough to meet the requirements, because the size of object is small, the environment is complex, the sample size is unbalanced and so on. Therefore, we design a simple but effective feature fusion method, called Double Context Information Enhancement Module. The model can fully learn the features of fasteners by expanding the receptive field of the feature map during the construction of the feature pyramid, and the feature expression ability of fasteners can be enhanced by learning the features of adjacent layers after the construction of the feature pyramid. Through experiments, we proved that DCEM can improve about 8.8AP compared with the original model, and at the same time, the effectiveness and advancement of DCEM are proved by comparing with SOTA.
Homography estimation is a crucial problem in computer vision, which aims to provide an optimal transformation matrix for aligning images captured from different viewpoints. Current methods extract shallow features fr...
Homography estimation is a crucial problem in computer vision, which aims to provide an optimal transformation matrix for aligning images captured from different viewpoints. Current methods extract shallow features from image pairs and introduce learnable mask modules to improve homography estimation performance. However, they struggle to capture long-term dependencies between features and comprehend the global structures of image features. A deep unsupervised homography learning framework is proposed in this paper, consisting of a weight-sharing feature extraction network and a homography estimation network based on the Transformer model. The former extracts the local features of images, while the latter learns the correlation between them and understands the global features of images, enabling the algorithm to better estimate the homography of unaligned images. Experimental results demonstrate that the proposed method outperforms the advanced methods for estimating homography matrices in the CA-Unsupervised dataset.
暂无评论