摘要
针对固体发动机电子计算机断层扫描(CT)图像数据具有伪影噪声的问题,和实际成型药柱燃面粗糙度大、退移计算难度大的问题,提出一种CT数据的快速最小距离函数(CTFMDF)法。固体火箭发动机CT图像中存在伪影噪声,采用非均值滤波(NLM)算法对CT图像进行去噪处理,采用Scharr算子对去伪影后的图像进行Canny边缘检测,提取装药燃面。最大类间方差(OTSU)算法将装药分离,并建立三维装药体数据模型,对燃面数据建立多个并行K-d树,快速检索出装药到燃面的最小距离。实验结果证明:对于不同的装药结构,可以完成任意燃去厚度时的燃面位置,且CT-FMDF法运行时间更短。基于CT实测数据,对于带有初始燃面缺陷的装药,可计算出燃烧时缺陷对燃面的影响。
A fast minimum distance function method for CT data(CT-FMDF)is proposed to address the artifact noise in CT image data of solid rocket motors,the large combustion surface roughness of actually formed grains,and the difficulty in transition calculation.The non-local means(NLM)filtering algorithm is used to denoise the CT images.Then,the Canny edge detection improved by the Scharr operator is used to extract the initial burning surface of the grains from the de-artifacted image.The maximum between-class variance algorithm(OTSU)separates the grains,establishing a three-dimensional volume data model of grains,and multiple parallel K-dimension trees for the combustion surface data.The minimum distance from the grains to the combustion surface is quickly retrieved.The experimental results show that for different grains structures,the position of a burning surface at any thickness can be achieved and that the algorithm has a shorter running time.For the grains with initial burning surface defects,the effect of the defects on the burning surface during combustion can be calculated based on the actual CT data.
作者
刘舜
卢洪义
张维维
章斌
杨禹成
桑豆豆
LIU Shun;LU Hongyi;ZHANG Weiwei;ZHANG Bin;YANG Yucheng;SANG Doudou(School of Aircraft Engineering,Nanchang Hangkong University,Nanchang 330063,China;Inner Mongolia Institute of Power Machinery,Hohhot 010010,China)
出处
《北京航空航天大学学报》
EI
CAS
CSCD
北大核心
2023年第11期3115-3123,共9页
Journal of Beijing University of Aeronautics and Astronautics
基金
江西省自然科学基金(20201BBE51002)
江西省研究生创新专项资金项目(YC2021-S685)。
