Research On Key Test Technology Of 3D Fragment Distribution Based On Binocular Vision

Fragmentation is one of the important damage elements of long-distance killing,the motion parameters of fragmentation test is an important step to evaluate its damage characteristics,fragmentation speed is an important parameter to measure the warhead power,but with the continuous development of various technologies,the fragmentation of motion parameters is not limited to the fragmentation velocity related research,more toward the direction of the overall motion process and distribution of fragments.Fragmentation in the process of movement accompanied by strong interference phenomenon,easy to damage the equipment,and fragmentation movement speed,wide distribution range,in the data acquisition and processing have a certain degree of difficulty.The test range of the existing test methods is affected by the system,which is generally used to test a small number of directional fragments.In view of these limitations,based on the advantages of binocular vision technology in solving three-dimensional coordinates,this paper studies the key technologies used to test the threedimensional distribution of fragments,and the research content is as follows:1.Research on calibration method of binocular camera: Aiming at the problems of large field of view,long target distance,small size of common calibration equipment,and difficult to cover field of view due to the constraints of frame structure,a new calibration method is proposed,in this method,the internal parameters of left and right cameras are calculated by using Zhang’s calibration method,and the external parameters of left and right cameras are solved by using DLT principle,and the structural parameters are obtained according to the external parameters;finally,within the field of view range of 4m*2.5m*6m,the calibration accuracy is verified by measuring the length of the target rod.The results show that the MAE and MRE of the target rod obtained by the method in this paper are 0.774 mm and 0.077%,the measuring accuracy is higher.2.Research on object recognition method of complex background fragment sequence image:Aiming at the problems of complex background,small target proportion and unclear gray distribution around the target,this paper firstly studies the background removal method of fragment image;secondly,the effect of several classical threshold segmentation methods on image segmentation is compared.After that,HBA-optimized Tsallis relative entropy algorithm is used to segment the image,and finally the recognition effect is verified.The experiment shows that under complex background conditions,the effective fragment recognition rate calculated by this method can reach 83.8%.3.Research on binocular vision matching method of fragment target: Aiming at the problem of low feature of fragments in the image and mutual occlusion of objects,the Bouguet stereo correction method is first used to achieve the alignment of left and right images;secondly,target area and candidate box size were used to distinguish and separate the target under the occlusion condition,and then the line-by-line matching method is used to match the target in the left and right images.Finally,experiments are designed to verify the results.The results show that under the condition of simple background,the method can successfully match all the targets in 22 groups of images.