航空激光增材制造零部件潜在缺陷检测

Potential defects detection of aviation laser additive manufacturing parts

传统的航空激光增材制造零部件潜在缺陷检测方法检测准确度低,图像特征查全率低。基于上述问题,提出一种基于图像识别技术的航空激光增材制造零部件潜在缺陷检测方法。采用红外成像技术进行航空激光增材制造零部件成像处理,提取航空激光增材制造零部件红外图像的缺陷区域特征点,对红外图像进行中值滤波降噪处理,利用扫描图像的纹理异常分布特性进行潜在缺陷的自适应定位检测,结合模板匹配和角点检测方法,实现对航空激光增材制造零部件潜在缺陷检测。仿真结果表明,采用该方法方法的图像特征提取查全率比传统方法提高了15%-20%,能够清晰检测到增材零件的潜在缺陷。说明进行航空激光增材制造零部件潜在缺陷检测的准确性较好,对缺陷部位定位的误差较小。

The traditional defects detection method of aviation laser additive manufacturing parts has low detection accuracy and low image feature recall.Based on the problems above,a method for detecting potential defects of aviation laser additive manufacturing parts based on image recognition technology is proposed.Use infrared imaging technology to perform imaging processing of aviation laser additive manufacturing parts.Extract feature points of defect areas in infrared images of aviation laser additive manufacturing parts.Perform median filtering and noise reduction processing of infrared images,and use abnormal texture distribution characteristics of scanned images to carry out adaptive positioning detection of potential defects.Combining with template matching and comer detection methods to realize potential defects detection of aviation laser additive manufacturing parts.The simulation results show that the recall rate of image feature extraction using this method is 15-20%higher than traditional method,and the potential defects of additive parts can be clearly detected,which shows that the accuracy of potential defects detection of aviation laser additive manufacturing parts is better,and the error of positioning defects is small.