针对毫米波雷达和视觉传感器融合算法在特征融合层面缺乏有效监督的问题,提出了一种引入激光雷达监督的多模态融合三维目标检测算法(Radar and Camera Fusion Based on Lidar Supervision,LRCFusion)。该算法首先分别提取视觉传感器、...针对毫米波雷达和视觉传感器融合算法在特征融合层面缺乏有效监督的问题,提出了一种引入激光雷达监督的多模态融合三维目标检测算法(Radar and Camera Fusion Based on Lidar Supervision,LRCFusion)。该算法首先分别提取视觉传感器、激光雷达和毫米波雷达各自的数据特征;接着使用知识蒸馏的方法,利用激光雷达特征作为教师模型监督毫米波雷达特征,以提升毫米波雷达特征的表达水平;然后引入注意力机制实现毫米波雷达和视觉特征融合,并采用基于点云的三维物体检测方法对融合的特征进行目标检测和3D锚框预测;最后,使用预测的3D锚框更新融合前的3D参考点。与基线算法进行比较,所提算法的平均精度提高1.2%,归一化检测得分提高1%。展开更多
Based on the characteristics of line structured light sensor, a speedy method for the calibration was established. With the coplanar reference target, the spacial pose between camera and optical plane can be calibrate...Based on the characteristics of line structured light sensor, a speedy method for the calibration was established. With the coplanar reference target, the spacial pose between camera and optical plane can be calibrated by using of the camera’s projective center and the light’s information in the camera’s image surface. Without striction to the movement of the coplanar reference target and assistant adjustment equipment, this calibration method can be implemented. This method has been used and decreased the cost of calibration equipment, simplified the calibration procedure, improved calibration efficiency. Using experiment, the sensor can attain relative accuracy about 0.5%, which indicates the rationality and effectivity of this method.展开更多
一当场，自我本地化系统为在有深入的 3D 里程碑的 3D 环境起作用的活动机器人被开发。机器人通过合并从 odometry 和单向性的照相机收集的信息的一个地图评估者递归地估计它的姿势。我们为这二个传感器造非线性的模型并且坚持说机器人...一当场，自我本地化系统为在有深入的 3D 里程碑的 3D 环境起作用的活动机器人被开发。机器人通过合并从 odometry 和单向性的照相机收集的信息的一个地图评估者递归地估计它的姿势。我们为这二个传感器造非线性的模型并且坚持说机器人运动和不精密的传感器大小的无常操作应该全部被嵌入并且追踪我们的系统。我们在一个概率的几何学观点和使用 unscented 变换描述无常框架宣传无常，它经历给定的非线性的功能。就我们的机器人的处理力量而言，图象特征在相应投射特征的附近被提取。另外，数据协会被统计距离评估。最后，一系列系统的实验被进行证明我们的系统的可靠、精确的性能。展开更多
This research is dedicated to develop a safety measurement for human-machine cooperative system, in which the machine region and the human region cannot be separated due to the overlap and the movement both from human...This research is dedicated to develop a safety measurement for human-machine cooperative system, in which the machine region and the human region cannot be separated due to the overlap and the movement both from human and from machines. Our proposal here is to automatically monitor the moving objects by image sensing/recognition method, such that the machine system can get enough information about the environment situation and about the production progress at any time, and therefore the machines can accordingly take some corresponding actions automatically to avoid hazard. For this purpose, two types of monitor systems are proposed. The first type is based on the omni directional vision sensor, and the second is based on the stereo vision sensor. Each type may be used alone or together with another type, depending on the safety system's requirements and the specific situation of the manufacture field to be monitored. In this paper, the description about these two types are given, and as for the special application of these image sensors into safety control, the construction of a hierarchy safety system is proposed.展开更多
文摘针对毫米波雷达和视觉传感器融合算法在特征融合层面缺乏有效监督的问题,提出了一种引入激光雷达监督的多模态融合三维目标检测算法(Radar and Camera Fusion Based on Lidar Supervision,LRCFusion)。该算法首先分别提取视觉传感器、激光雷达和毫米波雷达各自的数据特征;接着使用知识蒸馏的方法,利用激光雷达特征作为教师模型监督毫米波雷达特征,以提升毫米波雷达特征的表达水平;然后引入注意力机制实现毫米波雷达和视觉特征融合,并采用基于点云的三维物体检测方法对融合的特征进行目标检测和3D锚框预测;最后,使用预测的3D锚框更新融合前的3D参考点。与基线算法进行比较,所提算法的平均精度提高1.2%,归一化检测得分提高1%。
文摘Based on the characteristics of line structured light sensor, a speedy method for the calibration was established. With the coplanar reference target, the spacial pose between camera and optical plane can be calibrated by using of the camera’s projective center and the light’s information in the camera’s image surface. Without striction to the movement of the coplanar reference target and assistant adjustment equipment, this calibration method can be implemented. This method has been used and decreased the cost of calibration equipment, simplified the calibration procedure, improved calibration efficiency. Using experiment, the sensor can attain relative accuracy about 0.5%, which indicates the rationality and effectivity of this method.
基金Supported by National Natural Science Foundation of China(60605023,60775048)Specialized Research Fund for the Doctoral Program of Higher Education(20060141006)
文摘一当场，自我本地化系统为在有深入的 3D 里程碑的 3D 环境起作用的活动机器人被开发。机器人通过合并从 odometry 和单向性的照相机收集的信息的一个地图评估者递归地估计它的姿势。我们为这二个传感器造非线性的模型并且坚持说机器人运动和不精密的传感器大小的无常操作应该全部被嵌入并且追踪我们的系统。我们在一个概率的几何学观点和使用 unscented 变换描述无常框架宣传无常，它经历给定的非线性的功能。就我们的机器人的处理力量而言，图象特征在相应投射特征的附近被提取。另外，数据协会被统计距离评估。最后，一系列系统的实验被进行证明我们的系统的可靠、精确的性能。
文摘This research is dedicated to develop a safety measurement for human-machine cooperative system, in which the machine region and the human region cannot be separated due to the overlap and the movement both from human and from machines. Our proposal here is to automatically monitor the moving objects by image sensing/recognition method, such that the machine system can get enough information about the environment situation and about the production progress at any time, and therefore the machines can accordingly take some corresponding actions automatically to avoid hazard. For this purpose, two types of monitor systems are proposed. The first type is based on the omni directional vision sensor, and the second is based on the stereo vision sensor. Each type may be used alone or together with another type, depending on the safety system's requirements and the specific situation of the manufacture field to be monitored. In this paper, the description about these two types are given, and as for the special application of these image sensors into safety control, the construction of a hierarchy safety system is proposed.
