Research On Automatic Adjustment And Measurement Technology Of Dropping Attitude For Large Samples

The large-scale sample drop test is a drop test for large precision instruments,dangerous ammunition,fire control system and other large test samples.It is an important means to assess the safety and reliability of the test samples against impact.This method provides the basis for the impact resistance design of sample products and packaging.Aiming at the problem of insufficient drop test capability of large-scale samples,this paper presents a new system for the drop test of large scale samples.New system combines mechanical design technology,automatic control technology and computer vision processing technology to realize the automatic adjustment and measurement of the samples falling attitude.In this way,the instantaneous attitude angle of the sample falling to the ground is obtained.The working principle of the large-scale sample drop system and the whole structure of drop frame are studied.These research contents include grasping and releasing mechanism,angle adjusting mechanism and so on.In order to solve the problems of lifting large samples in the air and the difficulty of initial attitude adjustment and measurement,this paper proposes a method of feature extraction of Hough transform line of adaptive voting threshold to obtain the attitude angle of the sample.And further judging the difference between the angle and the preset attitude,the drop attitude can be adjusted automatically based on the motor position control algorithm.Through the research on the three-dimensional reconstruction method of high-speed photography binocular vision,the instantaneous attitude measurement of the large-scale test object upon landing is realized.The specific method is to analyze the imaging model of the high-speed camera and give the transformation relation of each coordinate system.And choose Zhang Zhengyou calibration method to realize camera calibration.Based on the study of SURF feature matching algorithm,the RANSAC algorithm is used to screen out mismatches to improve the matching accuracy problems of this algorithm.To solve the problem of invisibility in sparse matching reconstruction,proposing the algorithm SURF and seed region growth to achieve parallax propagation interpolation to solve the problem.Under the parallax continuity constraint,gray scale similarity constraint and uniqueness constraint,the dense parallax map is obtained,which makes the reconstruction visual.Furthermore,using the triangulation method to solve the three-dimensional coordinates of the samples,and the central axis method is used to obtain the attitude angle.The experimental setup was simulated to verify the feasibility of the proposed algorithm.The point cloud was used to reconstruct the samples and calculate the final attitude angle.