无人系统多模态视觉感知单元测试与评估方法的研究

Research on Test and Evaluation Method of Unmanned System Multi-Modal Vision Sensor Unit

提出一种通用的基于视觉传感器的无人系统多模态视觉感知平台的测试与评估的方法,研究了可见光波段和红外光波段下的动态测距实验与复杂光变测距实验.整个过程采用基于北斗卫星实时授时的时间同步技术和真值采集设备,输出带有时间戳的无人系统多模态感知单元的测量数据和真值数据,并保证了不同设备之间的时间同步误差小于10 ms.最后,对比测量数据和真值数据,给出测试设备的性能与功能的量化评估.实验结果表明,该测试评估方法有效地获取了真值,并进行了时间同步,满足实际的需求,具备推广性.

A universal method is proposed for testing and evaluating a multi-modal vision perception platform for unmanned systems based on vision sensors,and dynamic ranging experiments and complex optical variable ranging experiments under the visible and infrared light bands are studied.Time synchronization technology based on the real-time time service of Beidou satellite and accurate true value acquisition equipment are used in the test process,so that the time-stamped measurement data given by the unmanned system multi-modal vision perception unit and the time-stamped true values available for ranging performance evaluation are taken as the premise that the time synchronization errors between different devices are less than 10 ms.Finally,the measurement data and true value data are compared,and a quantitative evaluation of the performance and function of the test equipment are given.The experimental results show that the test and evaluation method effectively obtains the true value and synchronizes the time,which meets the actual needs and which can be popularized.