Research On A Network Planning Strategy And Experimental For Multi-vision Measurement Network

In large size industrial manufacturing,the development of multi-vision network measurement technology has become an effective means of 3-D profile measurement.But in the process of measurement,mostly to the measurement accuracy as the main consideration of the technology,which ignores the influence of coverage on the measuring system,and it will lead to complete morphology unable to reconstruct the measurement object,therefore,aiming at the requirements of high coverage and high accuracy in 3-D shape measurement,this paper proposes an intelligent network planning method considering the measurement coverage and 3-D uncertainty,researched the visual programming problem in 3-D profile measurement,and establish the corresponding measurement network planning algorithm.Firstly,the methods of 3-D vision measurement are analyzed and studied.The linear pinhole principle model of the camera and the principle model of the nonlinear projection are analyzed.In view of binocular stereo vision system,research the matrix transformation relationship between the three coordinate systems.A binocular stereo vision measurement system model and a multi-vision measurement network system model have been established,which provides a theoretical basis for the following chapters.Then,combining with the demand of multi-vision measurement network planning,the measurement network coverage and three-dimensional uncertainty are given respectively as the planning objectives.The discretization model of visual measurement network is designed,the decision variables of multi-vision measurement network are determined,the visibility and other constraints are considered,the mathematical model of multi-vision measurement network planning problem is established,and the multi-objective genetic algorithm is used to solve the problem of multi-vision measurement network planning.Design the genetic algorithm encoding and population processing method,fitness function and genetic operators,combined with the theoretical analysis,the propeller structure model is taken as an example to simulate the experiment,and a set of solutions that meet the planning requirements are obtained.The coverage of the measurement network can reach 99.72%,the three-dimensional uncertainty can converge to 0.0326 mm,the effectiveness and feasibility of the strategy are verified by a single vision multi station measurement experiment.Finally,the multi-vision measuring network is divided into multiple groups of binocular stereo vision measuring units,a binocular vision measurement unit as a node of measurement network system,and established the mathematical model of measuring network nodes pose parameters.The influence of the projection angle,the angle of the light axis and the baseline,the focal length and the baseline distance on the measurement accuracy of the system is analyzed.The optimum value range of the position and pose parameters of the network node is obtained.Finally,the accuracy of the theoretical analysis is verified by experiments.In this paper,based on the intelligence of vision measurement network,a multi vision measurement network planning strategy is designed.The research contents provide theoretical and technical support for multi vision measurement networking technology.