Convolutional neural network-based pose mapping estimation as an alternative to traditional hand-eye calibration.

The vision system is a crucial technology for realizing the automation and intelligence of industrial robots, and the accuracy of hand-eye calibration is crucial in determining the relationship between the camera and robot end. Parallel robots are widely used in automated assembly due to their high positioning accuracy and large carrying capacity, but traditional hand-eye calibration methods may not be applicable due to their limited motion range and resulting accuracy problems. To address this issue, we propose using a pose, nonlinear mapping estimation method to solve the hand-eye calibration problem and have constructed a 1-D pose estimation convolutional neural network (PECNN) with excellent performance, through experiments and discussions. The PECNN achieves an end-to-end mapping of the variation of the target object pose to the variation of the robot end pose. Our experiments have shown that the proposed hand-eye calibration method has high accuracy and can be applied to the automated assembly tasks of vision-guided parallel robots. Moreover, the method is also applicable to most parallel robots and tandem robots.

Research on registration error compensation of large-scale measurement field based on neural network.

The combination of large tooling size, environmental vibration, and equipment errors at the aircraft assembly site leads to errors in the enhanced reference system (ERS) point measurement information. ERS point errors directly reduce the accuracy of the assembly measurement field. This paper proposes ERS point error prediction and registration compensation based on the neural network to address this problem. First, the effects of equipment measurement errors and environmental vibration factors on the measurement field are studied. The ERS point error prediction model based on the neural network is established. On this basis, model evaluation is used to assess the prediction model of this paper. Then, a measurement field registration compensation model is constructed based on the neural network error results for ERS point compensation analysis. Finally, an experimental validation platform was built to predict the ERS point errors and compensate for the constructed measurement fields using the method in this paper. The experimental results show that, compared with the conventional method, the maximum registration errors in the X, Y, and Z directions are reduced from 0.0812, -0.0565, and -0.2810 to -0.0184, -0.0010, and 0.0022 mm, respectively, after compensation in this paper. The method proposed in this paper can not only predict the ERS point error state and provide a reference for designers but also guide the selection of appropriate ERS points when constructing the measurement field. The compensation method in this paper effectively reduces the measurement field registration error.

Improving pose estimation accuracy for large hole shaft structure assembly based on super-resolution.

Image resolution is crucial to visual measurement accuracy, but on the one hand, the cost of increasing the resolution of the acquisition device is prohibitive, and on the other hand, the resolution of the image inevitably decreases when photographing objects at a distance, which is particularly common in the assembly of large hole shaft structures for pose measurement. In this study, a deep learning-based method for super-resolution of large hole shaft images is proposed, including a super-resolution dataset for hole shaft images and a new deep learning super-resolution network structure, which is designed to enhance the perception of edge information in images through the core structure and improve efficiency while improving the effect of image super-resolution. A series of experiments have proven that the method is highly accurate and efficient and can be applied to the automatic assembly of large hole shaft structures.

A measurement point planning method based on lidar automatic measurement technology.

The automated measurement mode in the lidar measurement system (LMS) has advantages unmatched by other measurement equipment in measuring the surface of large components. Before starting the automatic measurement mode, it is necessary to plan the measurement guide points for the measurement area. The rationality of the measurement points planning will directly affect the quality of the measurement data and the measurement efficiency. This paper proposes a planning method for measurement points on the outer surface of components based on lidar automatic measurement technology. First, the geometric features to be measured are discretized into spatial point cloud data. Second, the edge points of the point clouds are extracted and indented to meet the measurement requirements, which improves the measurement accuracy of the edge areas. Finally, by planning the path of the measurement points, the laser beam of the lidar can traverse the points of the measurement features with the shortest paths. Through the analysis of two cases, the method proposed in this paper will provide a huge advantage for the LMS: (1) The edge points of all features can be identified and indented in a short time to ensure the measurement accuracy of the edge areas of each measurement feature. (2) Through the measurement path planning, the repetitive measurement path of the lidar can be significantly reduced, improving the measurement efficiency. The method proposed in this paper has important guiding significance for the subsequent measurement station planning and constructing large-scale spatial measurement fields of the LMS.

Surface reconstruction based on CAD model driven priori templates.

In this paper, a method of reconstructing 3D (three dimensional) models from the original scanned point cloud using priori templates is proposed. Different from previous reconstruction methods that triangulate and fit the original scanning point cloud directly, we construct a priori template based on the CAD (computer aided design) model and guide the reconstruction of the original scanning point cloud with the priori template. Given a CAD model, the basic geometric elements are used as the basic units to extract the set elements of 3D shapes. Then the geometric elements are meshed, and the normal vectors at the mesh nodes are extracted. The corresponding point cloud data of each basic element are extracted from the original point cloud. The point cloud data near the normal of the guide point are searched, and the Gaussian weighted average value of the searched point represents the actual geometric parameters of the part at the guide point. Finally, the geometric elements of the basic unit are reconstructed locally by Non-Uniform Rational B-Splines surface fitting, and the complete reconstruction model is obtained by integrating the local reconstruction. Experiments show that our method can solve the problems of high quality reconstruction, sharp feature preservation, and detail recovery in surface reconstruction.

[Finite element stress analysis of all-ceramic crown of the upper central incisor in differential thickness].

OBJECTIVE: To investigate the stress distributions under load in an all-ceramic crown of the upper central incisor in differential thickness. METHODS: The 3-dimensional finite element model of all-ceramic crown of the upper central incisor in differential thickness was applied with differential loads (100, 150, 200 N). The stress values and distributions of all-ceramic crown were calculated and expressed. RESULTS: The tendency of stress distributions in all-ceramic crown of differential thickness and loads was similar. The maximal stress intensity value was located in the loading site and the cervical region. As the thickness of all-ceramic crown increased, the stress concentration reduced. When the load increased, the stress values increased, especially in the loading site and the cervical region. CONCLUSION: The increasing of the thickness of all-ceramic crown will decrease stress concentration. Because stress concentration is mainly in cervical region and loading site, these regions should be cared in fabrication of all-ceramic crown.