Research On Calibration And Optimization Of Large Field Based On Trinocular Stereo Vision

In recent years,with the development of China's large-scale equipment manufacturing industry,the requirement of visual inspection has gradually developed from small view field inspection to large view field inspection.The first step for large view field inspection is camera calibration.And camera calibration needs to make a high-precision calibration template which approaches to the surface size of measurement object.However,it is extremely difficult to make the calibration template matching with the object which is measured by large view field inspection.So,the traditional calibration method is difficult to meet its high-quality testing requirements.The domestic and foreign scholars have mainly concentrated on the binocular vision system in the analysis of optimal structural parameters for a long time,which can not meet the calibration requirements of large view field measurement.So,the large view field calibration based on the trinocular stereo vision and optimization method research are conducted in this paper for such kind of problem.The accuracy is improved in this paper by selecting the optimum structural parameters of the system and reducing the errors produced by the camera.Finally,a scalibration scheme for large view field based on trinocular stereo vision is proposed and a calibration optimization method is studied based on the calibration scheme.This research provides some technical support for improving the detection precision of large equipment manufacturing industry.First,a separation-type calibration algorithm is proposed for the constraint condition of object which is measured by large view field.And,the separationtype calibration principle is analyzed based on this algorithm,then the Rodrigue matrix is used to work out transformation matrix.Also,the calibration thought based on trinocular stereo vision system is put forward in accordance with the restriction of binocular vision system in detecting the field angle scope of object.Then,the putting place of three cameras is determined in trinocular stereo vision system in accordance with the measurement mechanism of large view field.Finally,a separation-type calibration scheme for large view field based on trinocularstereo vision is proposed.Also,the validity of calibration method proposed in this paper is verified by experiment.Then,the calibration optimization method based on large field calibration is studied because the research is required to further improve the accuracy of threedimensional reconstruction result.In this paper,the mathematical model of trinocular stereo vision system is established to obtain the relationship between three-dimensional coordinates and structural parameters of the camera position.Based on this,the relational graph between error transfer coefficient and structural parameters is drawn on the basis of error transfer coefficient minimization criterion.Then,the effect of structural parameter on accuracy is comprehensively analyzed based on the relational graph so as to obtain the optimal range of the structural parameters.The results show that the optimized large FOV separation calibration method can well achieve large field calibration.