叶型孔图像拼接技术实验研究  

Experimental Study on Image Mosaic Technology of Blade-shaped Hole

作　　者：张超 盛波 毕超 徐微雨 ZHANG Chao;SHENG Bo;BI Chao(AVIC Beijing Precision Engineering Institute for Aircraft Industry,Beijing 100076;AECC South Industry Company Limited,Zhuzhou 412002)

机构地区：[1]航空工业北京航空精密机械研究所,北京100076 [2]中国航发南方工业有限公司,株洲412002

出　　处：《航空精密制造技术》2023年第2期13-16,12,共5页Aviation Precision Manufacturing Technology

摘　　要：为了提高叶型孔图像拼接的精度,开展了叶型孔图像拼接实验研究,使用Harris、SIFT、SURF3种特征点检测算法对同一张叶型孔图像进行特征点检测,通过SSIM和轮廓度两种评价方法评价了BruteForce+KNN、Br uteForce+CrossCheck和K-D tree+KNN3种特征点匹配算法得到的拼接图像,实验结果表明选用Harris进行特征点检测,BruteForce+KNN进行特征点匹配,利用RANSAC求变换矩阵,得到像素点加权融合重合部分的拼接图像精度最高,为叶型孔非接触测量奠定基础。In order to improve the accuracy of leaf hole image mosaic,experimental research on leaf hole image mosaic is carried out.Harris,SIFT and SURF feature point detection algorithms are used to detect the feature points of the same leaf-shaped hole image.The mosaic images obtained by BruteForce+KNN,BruteForce+CrossCheck and K-D tree+KNN feature point matching algorithms were evaluated by two evaluation methods:SSIM and contour.The experimental results show that Harris is used for feature point detection,BruteForce+KNN is used for feature point matching,and RANSAC is used to find the transformation matrix to obtain the highest accuracy of the mosaic image of pixels weighted fusion overlap part,which lays a foundation for the subsequent leaf hole project.

关 键 词：航空发动机 叶型孔 图像拼接 图像评价 

分 类 号：TH72[机械工程—仪器科学与技术]