视觉测量数据融合技术研究

结构光测距技术的主要缺点是无法获取摄像机看不到或光源照不到的点的数据。为此，采用冗余多传感器构成视觉坐标测量系统。传感器输出各自测得的目标参数和状态参数到数据融合中心，特殊设计的惯量计算方法抽提出各路信号的特征参数。利用特征参数阈值进行数据特征融合后，再进行加权融合计算，使融合后的最终测量结果达到最优。实验证明数据融合技术明显改善测量系统性能，对复杂形貌曲面测量有重要意义。

Investigation on the Data Merger Technique For Vision Measuring System

The main shortcoming of structural light telemeter is that the sightless points for the sensors or the photosource can not be obtained. So, a vision measuring system is constituted of more redundancy sensors. Every sensor obtains its target parameters and status parameters and outputs to the computer. Special method of inertia calculation refines on every feature parameters. After merging the data feature by use of threshold, we merge the data with weighing calculation to obtain optimized results. The experiment proved that the data merger will improve the system performance obviously, and this is important to measure a complicated quadric surface.