Shenyang Institute of Automation,Chinese Academy Of Sciences
Not a member yet
23582 research outputs found
Sort by
MSL3D: 3D object detection from monocular, stereo and point cloud for autonomous driving
No full textIn this paper, we propose a novel deep architecture by combining multiple sensors for 3D object detection, named MSL3D. While recently LiDAR-Camera methods introduce additional semantic cues, working with fewer false detections, there is still a performance gap compared LiDAR-only methods. We argue that this gap is caused for two reasons: 1) the 3D spherical receptive fields of the set abstraction of the point clouds are not aligned with the 2D pixel-level receptive fields of the image. 2) the premature introduction of image information makes it is difficult to apply data augmentation both LiDAR and image synchronously. For the first problem, we extend 3D set abstraction to a 2D set abstraction that can transform the 2D image features to the 3D sphere to unify the receptive field of multi-modal data. For the second problem, we design a novel two-stage 3D detection framework that employs the LiDAR-only backbone in the first stage to estimate high-recall and high-quality proposals and then integrates the image and point clouds information for box refinement and confidence prediction. Besides, we add two auxiliary networks to effectively learn image features and point cloud features when using different multi-modal data augmentation strategies synchronously. Moreover, we design a consistency-structure generator using stereo images to determine whether any of a point in the 3D space belongs to the contour of the object, thereby supplementing the sparse point cloud information. Extensive experiments on the popular KITTI 3D objects detection dataset show that our proposed MSL3D achieves better performance comparing with other LiDAR-Only or LiDAR-Camera fusion approaches.</p
Influence of Autonomous Sailboat Dual-Wing Sail Interaction on Lift Coefficients
No full textTo analyze the influence of the chord length ratio and angle of attack on lift coefficients and explore the interaction mechanism between the two, we established a calculation model of the pressure distribution coefficient on the airfoil surface and lift coefficient of a dual-wing sail on the basis of the vortex panel method. Computational fluid dynamics was used in auxiliary calculation and analysis. Results revealed a reciprocal interference between the front-wing and rear-wing sails. The total lift coefficient of the dual-sail increased with an increase in the front sail chord length. The lift coefficient of the rear sail decreased with an increase in the front sail chord length or angle of attack. The front sail wake affected the pressure distribution on the upper and lower surfaces of the rear sail leading edge.</p
IL-MCAM: An interactive learning and multi-channel attention mechanism-based weakly supervised colorectal histopathology image classification approach
No full textIn recent years, colorectal cancer has become one of the most significant diseases that endanger human health. Deep learning methods are increasingly important for the classification of colorectal histopathology images. However, existing approaches focus more on end-to-end automatic classification using computers rather than human-computer interaction. In this paper, we propose an IL-MCAM framework. It is based on attention mechanisms and interactive learning. The proposed IL-MCAM framework includes two stages: automatic learning (AL) and interactivity learning (IL). In the AL stage, a multi-channel attention mechanism model containing three different attention mechanism channels and convolutional neural networks is used to extract multichannel features for classification. In the IL stage, the proposed IL-MCAM framework continuously adds misclassified images to the training set in an interactive approach, which improves the classification ability of the MCAM model. We carried out a comparison experiment on our dataset and an extended experiment on the HENCT-CRC-100K dataset to verify the performance of the proposed IL-MCAM framework, achieving classification accuracies of 98.98% and 99.77%, respectively. In addition, we conducted an ablation experiment and an interchangeability experiment to verify the ability and interchangeability of the three channels. The experimental results show that the proposed IL-MCAM framework has excellent performance in the colorectal histopathological image classification tasks
Study on Thermo Mechanical Coupling of the Hybrid Additive and Subtractive Manufacturing Based on Finite Element Model
No full text采用高斯移动热源、重启动分析、预定义场以及J-C本构模型等方法，探究了增减材复合制造过程中温度场与应力场的耦合作用，实现了增减材复合制造过程有限元模型的构建并通过实验验证了模型的准确性。研究结果表明，采用高斯线热源方法在降低计算量的同时，还可以保证模型的准确性，而重启动分析与预定义场实现了增减材制造过程工艺的良好衔接；随着温度的降低，应力形式由热应力转变为加工应力，工件表面由残余拉应力转变为残余压应力，表明在增材制造过程中激光加热产生的高温对工件表面应力影响较大；在增材工艺与减材工艺之间的冷却时间延长后，工件表面的残余应力由正值转化为负值，这说明减材工艺的介入时机对工件表面的应力状态有较大影响，选择合适的减材加工温度，有利于降低残余应力对工件的影响。</p
基于BP神经网络的旋回式破碎机故障诊断研究
No full text矿石破碎加工在采矿生产工艺中是一个极其重要的环节，由于破碎机等大中型设备在实际使用的过程中存在冲击强烈、磨损严重等缺点和不足，且存在各种类型故障和设备部件运动的高复杂性，因此，在实际作业中始终面临着设备部件维修周期长、故障发病率高的特点。以旋回式破碎机为实例，利用各类传感器实时采集数据，基于BP神经网络构建故障诊断模型，选取4大故障类型，包括平行轴油温异常、轴承磨损、平行轴缺油和偏心套故障，同时将轴承转速、润滑油油压、回流油温和轴承振动频率作为4大故障特征参数，使用样本数据对BP神经网络故障诊断模型进行训练和优化，并使用测试数据验证了旋回式破碎机故障诊断模型的准确性及可行性。</p
Hydrodynamic performance of propeller in cavitating conditions under different ambient pressures
No full text为研究水面快速游艇、水面大型舰船和水下机器人的高负荷螺旋桨，本文采用MAU4-40小型桨，通过改变环境压力，基于两相流和空化模型预报了3种载体螺旋桨空泡初生的临界转速，空泡发展形状和空泡生成后的螺旋桨水动力性能曲线。并将水下机器人携带纵斜桨自航空泡问题进行研究。结果表明：理论预报空泡发生的临界转速小于数值模拟的临界转速；3种工况对应不同的临界转速；当空泡覆盖整个叶背区域，推力系数最大下降28.5%;纵斜能降低空泡发生，并使空泡位置移向随边；水下机器人伴流对螺旋桨空泡略有影响。</p
Thrust Prediction Method of Jet Propulsion for Transformer Internal Inspection Robot
No full textRobot technology has been successfully applied to the internal inspection of large power transformers. In order to further improve the level of robot intelligent control and realize robot motion control accuracy of the floating state, this paper analyzes the factors that affect the jet thrust of the robot propulsion system on the developed spherical transformer internal inspection robot platform based on jet propulsion, and builds an experimental platform. Combining experimental data to establish jet thrust equation. On this basis, the working state of the robot power system is further analyzed, the equivalent discharge model of the robot battery is established, the battery state parameters are introduced into the jet thrust equation. Finally, a mathematical model of jet thrust prediction is established that comprehensively considers the robot power system, control system, and environmental parameters. The establishment of this model lays a good theoretical foundation for the establishment of the robot's precise dynamics model, fine motion control and intelligent control algorithm
Dynamic Analysis and Vibration Reduction of Articulated Silicone Gel Column With Varying Geometry
No full textTo improve the vibration reduction effect in low-frequency band of dynamic vibration absorber (DVA), a novel type of articulated silicone gel column (SGC) is introduced in the design of the tuned dynamic vibration absorber. The nonlinear variation of frequency of SGC with varying geometry is obtained by both finite element simulation and experiments. The most sensitive mode is located, which has a wider frequency range by varying the geometry. The polynomial fitting is used to describe nonlinear relation between frequency and geometry. By tuning the geometry, the equivalent stiffness and then resonance frequencies can be manipulated to behave as an active vibration absorber. The vibration reduction experiment of SGC vibration absorbers is investigated. It is found that SGC has better vibration reduction effect in the low-frequency band. The experimental results in the current design demonstrate that the vibration reduction effect can reach 94.03% when tuning SGC to the first-order main resonance. The dimensions and material parameters of SGC should be altered for the specific frequency range and vibration strength
Optimization of Engine Bay Segment Measurement Model Based on Point Cloud Denoising Algorithm
No full text由于激光测量得到的航天发动机舱段点云中通常包含噪声,为提高后续路径规划中模型的精确度,提出一种针对噪声的点云去噪算法。首先,根据张量投票算法,通过点的张量矩阵得到扩散张量;其次,通过扩散张量设计各向异性扩散滤波在不同方向的速率,实现点云的大尺度噪声去噪;最后,通过点云的半径和标准差来实现双边滤波的自适应调参,进一步实现小尺度噪声去噪。对本算法进行了对比实验验证,结果表明该算法和传统算法相比,在有效剔除噪声点的同时,更好地保持了点云的几何特征,是一种高效的去噪算法。</p
基于面元模型的非刚性动态场景重建方法研究
No full text场景目标几何变形和运动的同时表述是非刚性动态场景重建的一个关键问题,出现遮挡、开—合拓扑变化、快速运动等情况都可能导致场景重建失败。针对这一问题,研究提出了一种基于面元模型的非刚性动态场景重建方法,利用基于权值算法的深度相机点到平面迭代最近点（Iterative Closest Point,ICP）算法改进了已有系统非刚性变形场估计中ICP能量函数的求解过程。与基于SDF模型的重建方法相比较,基于面元模型的方法支持在线高效更新,占用计算资源更少,更加轻量化。在VolumeDeform数据集上的对比实验表明,改进后的方法提高了系统的鲁棒性,重建的模型更加完整。</p