Optics and Precision Engineering, Volume. 16, Issue 11, 2151(2008)
Application of real-time feature level multi-sensor data fusion system to photoelectric theodolite
In order to satisfy the requirement of real-time tracing for photoelectric theodolite with multi imaging systems and to guarantee tracing precision and reliability,a new servo-control system according to Multiple Instruction Multiple Data Stream(MIMD) structure was proposed.The predictive filtering and data fusion of theodolite was derived theoretically.The results show that solution of data fusion needs a vast amount of numerical calculations.In order to satisfy the calculation requirement,a special parallel calculation system with three arithmetic units,two DSPs and a single FPGA was developed.The time outgoing of real-time calculation for each calculation unit was analysed.The simulation results show that the system can achieve 1.25 ms operation time in real-time for the theodolite with three imaging systems,while the CCD sampling frequency is 50 Hz and sampling frequency of photoelectric encoder is 800 Hz.
Get Citation
Copy Citation Text
CHENG Guo-ying, ZHANG Shu-mei, CAI Sheng, GAO Hui-bin, QIAO Yan-feng. Application of real-time feature level multi-sensor data fusion system to photoelectric theodolite[J]. Optics and Precision Engineering, 2008, 16(11): 2151
Category:
Received: Jul. 24, 2008
Accepted: --
Published Online: Feb. 28, 2010
The Author Email: Guo-ying CHENG (chenggy1980@163.com)
CSTR:32186.14.