面向发动机喷管测量的多目拼接算法

Multi-view stitching algorithm for engine nozzle measurements

为测量发动机矢量喷管的运动姿态及运动参数,在双目视觉测量的基础上,将多个双目视觉单元间的数据进行拼接和融合。为提高双目视觉测量单元间的拼接精度,提出一种随机法和穷举法相结合的拼接算法,并给出了算法的理论推导过程和仿真实验结果数据。实验结果表明:本文算法具有不添加损失函数进行约束的优点,可以很好地解决稀疏公共点数据拼接困难的问题,且具有很高的测量精度,这对于发动机设计和研发具有重要的理论意义和实用价值。

The data between multiple binocular vision units were stitched and fused on the basis of binocu⁃lar vision measurement to measure the motion attitude and motion parameters of the engine vector nozzle.In order to improve the stitching accuracy between binocular vision measurement units,a splicing algorithm combining random method and exhaustive method was proposed,and the theoretical derivation process of the algorithm and the data of simulation experiment results were given.The experimental results show that the algorithm in this paper has the advantage of not adding a loss function to constrain,it can solve the diffi⁃culty of splicing sparse common point data,and has a high measurement accuracy,which is of great theoreti⁃cal significance and practical value for engine design and development.