Research On Pose Measuring Of Helmet Based On Grid Structured Light

Aiming at the problem that some pose measurement systems will increase the weight of the helmet,this paper designs a helmet pose measuring system based on grid structured light.This system obtains 3D point cloud of the helmet using grid structured light and combines the point cloud registration algorithm to complete the helmet pose measurement.The system does not need to install any hardware facilities on the helmet,which reduces the weight of the helmet and reduces the pilot’s flight burden.The main research contents of this paper are as follows:1.A helmet pose measuring system based on grid structured light is designed.A three-dimensional measurement mathematical model of grid structured light is constructed,and a calibration method of grid structured light is proposed.The equation of spatial light is obtained by fitting the coplanarity between the corner point of structured light and the target.The camera and structure light projector are assembled,the 3D measurement system is calibrated,and the parameters of the camera and structured light are obtained.2.The image processing technology of grid structured light is studied.The center of structured light is extracted by Zhang’s fast parallel refinement method,and the corner points of grid structured light are roughly extracted by Shi-Tomasi algorithm.A corner topological relationship confirmation algorithm for grid structured light is proposed.According to the light strip sorting results and the corner light strip spacing,an ordered corner array is guaranteed to be output.Through experiments,the accuracy of grid structure light corner point extraction is verified,and the three-dimensional point cloud is achieved based on the calibration results.3.The pose measurement algorithm of point cloud registration is studied.According to the density and quantity characteristics of point cloud,a point cloud registration pose measurement algorithm combining SAC-IA coarse registration and improved ICP fine registration is studied.The SAC-IA rough registration algorithm is used to complete the preliminary registration of point clouds.An improved ICP fine registration algorithm is proposed to improve the accuracy of point cloud registration in helmet pose measurement.The registration accuracy,registration efficiency and final pose measurement accuracy are verified by experiments.