基于图像预处理的发动机叶片掉块损伤自动测量技术

Automatic Measurement Technology of Engine Blade Damage Based on Image Processing

关键词:  图像处理  航空发动机叶片  损伤识别  自动测量  特征提取

Key words:  Image processing  Aeroengine blades  Damage identification  Automatic measurement  Feature extraction

摘要:

针对当前孔探仪检测飞机发动机叶片缺乏有效自动测量缺陷尺寸的问题，从处理孔探图像的角度进行研究，提出了一套适用于叶片损伤的图像预处理方式，并运用基于特征点的自动测量方法实现了发动机叶片损伤的识别测量。首先，在预处理阶段运用直方图均衡化、Grab Cut分割、空间域增强等算法实现了图像去噪、改善局部高光、简化预处理一般步骤等。其次，建立图像坐标扫描叶片轮廓线获取位置信息，利用曲线斜率特点拟合叶片边缘直线，提取损伤的三个特征点用以描述获取损伤尺寸。最后，以某型发动机叶片的损伤为例，运用此预处理方法提取到了叶片边缘曲线，测量结果与实际测量结果基本相符，验证了该算法识别测量的可靠性，为航空发动机叶片的损伤测量提供了新思路。

Abstract:

In view of the lack of effective automatic measurement of defect size in aircraft engine blade detection by hole detector， a set of image preprocessing method suitable for blade damage is proposed from the perspective of hole detection image processing， and the identification and measurement of engine blade damage is realized by using the automatic measurement method based on feature points. Firstly， in the preprocessing stage， histogram equalization， grab cut segmentation， spatial domain enhancement and other algorithms are used to realize image denoising， improve local highlights， simplify the general steps of preprocessing and so on. Secondly， the image coordinates are established to scan the blade contour to obtain the position information， the curve slope characteristics are used to fit the blade edge line， and three feature points of the damage are extracted to describe and obtain the damage size. Finally， taking the damage of an engine blade as an example， the blade edge curve is extracted by using this preprocessing method， and the measurement results are basically consistent with the actual measurement results， which verifies the reliability of the identification and measurement of the algorithm， and provides a new way for the damage measurement of aeroengine blades.