Research On Dynamic Coordinate Measurement Method Based On Orthogonal Cylindrical Imaging

Promoting the industrial processing manufacturing to intelligent manufacturing's transformation and upgrading is one of development trend of the future industry.As typical application of intelligent manufacturing,intelligent assembly and process monitoring require to realize the real-time measurement of multiple targets.Existing dynamic coordinate measuring methods,such as: laser tracking total-station measurement,photoelectric scanning network measurement and large-scale stereo vision measurement,have the respective limitations of signal target,nonsynchronous measurement and low measurement frequency during dynamic coordinate measurement.So it is necessary to seek a new method of high performance dynamic coordinate measurement to meet the continuous development of industrial demand.With respect to the limitations of existing dynamic coordinate measurement methods,this paper starts exploring from sensor level,and disassembles the function of two dimensional angle measurement that is provided by array image sensor using in vision measurement.Combining two orthogonal linear CCDs(Charge Coupled Device)with the orthogonal cylindrical imaging system,we realize the two dimensional angle measurement.Then three dimentional coordinate can be obtained by multi-angle intersection.The new measurement method guarantees the posibility of multi-target,high-frequency,synchronous measurement in principle.The orthogonal cylindrical imaging measurement system is treated as research object.Based on the error analysis of the proposed system,we have developed the research successively from three aspects: camera calibration,dynamic performance improvement and multi-target tracking.The main problem,solution and concrete implementation corresponding to the three aspects are detailly illuminated and a relatively complete dynamic coordinate measurement method system is established.Both simulative and real experiments are used to verify the proposed methods.Experimental results show that the orthogonal cylindrical imaging measurement system has good measurement accuracy and dynamic performance,and can accomplish the multi-target real-time tracking.The main research work of this paper is as follows:1.With respect to the proposed orthogonal cylindrical imaging measurement system,the principle of coordinate measurement based on multi-angle intersection is studied.Angle measurement model is established and coordinate reconstruction method is introduced.Then the implementation of the measurement system is provided in the lay of application.2.The main error sources in the proposed measurement system is analyzed.The error factors coming from single camera is analyzed so as to find a precise camera calibration method and the layout of multi-camera is analyzed so as to study the transfer relationship of coordinate error.The integration of both can improve the measurement accuracy of the proposd system.3.With respect to the limitation of traditional camera calibration methods owning to the complex lens distortion,a precise camera calbibration method based on radial basis function mapping is proposed.This method obeys the essence of camera calibration that is to recover the projection mapping of imaging system,and provides an accuracy guarantee for the measurement system.4.Given the random error during dynamic measurement,Kalman filter algorithm is adopted to promote the dynamic performance of measurement system.The dynamic coordinate measurement methods based on Kalman filter and extended Kalman filter are studied.5.In order to solve the problem of multi-target recognition that is a common problem for the cylindrical imaging system,a multi-target tracking and recognizing method based on extended Kalman prediction is proposed to realize the matching and recognition of multiple targets and their image points.Then considering the possibility of spot intersection and gross error,the corresponding solution is expounded.