基于高频增强和区域标定的涡轮叶片缺陷定位

Turbine Blade Defects Positioning Based on High Frequency Enhancement and Regional Calibration

提出一种基于高频增强和区域标定的涡轮叶片缺陷定位检测算法,设计涡轮叶片缺陷定位检测工作原理,采用高频滤波的方法来锐化叶片图像,采用预处理的办法来初步控制过度分割现象,由灰度图像生成梯度图像,构建一个噪声抑制算子,解决过度分割问题,进行了涡轮叶片的缺陷特征模型构建和缺陷信息计算,采用高频增强算法和区域标记的方法,减少由于噪声和纹理细节产生的伪边界,提高缺陷定位性能。仿真实验表明,采用该算法能有效提高涡轮叶片的缺陷定位检测性能,突出损伤边界,较好地抑制噪声。具有较高的工程应用价值。

A turbine blade defects positioning and regional calibration of the detection algorithm is proposed based on en?hanced high frequency, the design principle of turbine blade detection defect positioning method is taken, using high fre?quency filter to sharpen the blade image, using pretreatment methods to preliminary control over segmentation phenomenon generated by the gradient image, gray image, build a noise suppression operator, solution the over segmentation problem of feature model, the defects of turbine blade construction and defect information calculation, using high frequency enhance?ment algorithm and region labeling, due to the pseudo boundary noise and texture details generated defects, improve the po?sitioning performance. Simulation results show that, the algorithm can effectively improve the detection performance of tur?bine blade defects positioning, highlight the damage boundary, and restrain the noise. It has a higher application value in engineering.