Sparse bundle adjustment(SBA) is a key but time-and memory-consuming step in three-dimensional(3 D) reconstruction. In this paper, we propose a 3 D point-based distributed SBA algorithm(DSBA) to improve the speed and ...Sparse bundle adjustment(SBA) is a key but time-and memory-consuming step in three-dimensional(3 D) reconstruction. In this paper, we propose a 3 D point-based distributed SBA algorithm(DSBA) to improve the speed and scalability of SBA. The algorithm uses an asynchronously distributed sparse bundle adjustment(A-DSBA)to overlap data communication with equation computation. Compared with the synchronous DSBA mechanism(SDSBA), A-DSBA reduces the running time by 46%. The experimental results on several 3 D reconstruction datasets reveal that our distributed algorithm running on eight nodes is up to five times faster than that of the stand-alone parallel SBA. Furthermore, the speedup of the proposed algorithm(running on eight nodes with 48 cores) is up to41 times that of the serial SBA(running on a single node).展开更多
Background Currently,laser tracker is the primary instrument used to carry out three-dimensional position measurement in accelerator alignment.Theoretically,three-dimensional measuring data processed by three-dimensio...Background Currently,laser tracker is the primary instrument used to carry out three-dimensional position measurement in accelerator alignment.Theoretically,three-dimensional measuring data processed by three-dimensional adjustment are more rigorous,however,error accumulation is found in practice.Purpose In order to control error accumulation and further improve the measurement accuracy of accelerator alignment,this research introduces the laser alignment system into the activity of measurement and data processing.Methods A measurement scheme combining laser tracker and laser alignment system is proposed.To construct the constraint condition,the offset values from the measuring points to the laser straight-line datum were used.To carry out the three-dimensional adjustment with offset constraint,the laser tracker observations were used.Results A three-dimensional adjustment function model of laser tracker observations is given.The construction method of the constraint equation is researched,and the calculation formulas of the three-dimensional adjustment with offset constraint are derived.A 200 m linac tunnel control network is designed,using simulation measurement method,the measuring data of laser tracker and the offset values from the measuring points to the laser straight-line datum were generated.The simulated data are calculated by the method given in this paper and the result is analyzed.Conclusion Simulation result shows introducing the laser alignment system into laser tracker measurement and applying the three-dimensional adjustment with offset constraint can effectively suppress the error accumulation caused by long distance move station measurement.展开更多
基金Project supported by the National Natural Science Foundation of China(Nos.U1435219,U1435222,and 61572515)the National Key R&D Program of China(No.2016YFB0200401)the Major Research Plan of the National Key R&D Program of China(No.2016YFC0901600)
文摘Sparse bundle adjustment(SBA) is a key but time-and memory-consuming step in three-dimensional(3 D) reconstruction. In this paper, we propose a 3 D point-based distributed SBA algorithm(DSBA) to improve the speed and scalability of SBA. The algorithm uses an asynchronously distributed sparse bundle adjustment(A-DSBA)to overlap data communication with equation computation. Compared with the synchronous DSBA mechanism(SDSBA), A-DSBA reduces the running time by 46%. The experimental results on several 3 D reconstruction datasets reveal that our distributed algorithm running on eight nodes is up to five times faster than that of the stand-alone parallel SBA. Furthermore, the speedup of the proposed algorithm(running on eight nodes with 48 cores) is up to41 times that of the serial SBA(running on a single node).
基金Funding was provided by Young Scientists Fund(Grant no.12075264).
文摘Background Currently,laser tracker is the primary instrument used to carry out three-dimensional position measurement in accelerator alignment.Theoretically,three-dimensional measuring data processed by three-dimensional adjustment are more rigorous,however,error accumulation is found in practice.Purpose In order to control error accumulation and further improve the measurement accuracy of accelerator alignment,this research introduces the laser alignment system into the activity of measurement and data processing.Methods A measurement scheme combining laser tracker and laser alignment system is proposed.To construct the constraint condition,the offset values from the measuring points to the laser straight-line datum were used.To carry out the three-dimensional adjustment with offset constraint,the laser tracker observations were used.Results A three-dimensional adjustment function model of laser tracker observations is given.The construction method of the constraint equation is researched,and the calculation formulas of the three-dimensional adjustment with offset constraint are derived.A 200 m linac tunnel control network is designed,using simulation measurement method,the measuring data of laser tracker and the offset values from the measuring points to the laser straight-line datum were generated.The simulated data are calculated by the method given in this paper and the result is analyzed.Conclusion Simulation result shows introducing the laser alignment system into laser tracker measurement and applying the three-dimensional adjustment with offset constraint can effectively suppress the error accumulation caused by long distance move station measurement.
