图像识别技术及其在机械零件无损检测中的应用

无损检测法是一种常用的故障诊断技术,故障诊断从本质上来讲就是模式识别问题,而模式识别又可以狭义地理解为图像识别。从介绍图像、图像识别、图像识别过程和图像识别系统的基本概念着手,就几种常用图像识别方法的原理和特点进行比较,给出了CCD图像获取系统的组成。最后,结合发动机曲轴的一种自动磁粉探伤系统实例,对系统的图像处理和识别流程进行详细的讨论,并针对一般无损检测系统难以满足曲轴的检测要求和精度要求的状况,提出经过改进的一种适用于曲轴的整体无损检测系统。该系统有助于高效和完整地获取整个曲轴的图像,提高图像信息的质量,从而提高发动机曲轴表面缺陷检测的准确性和可靠性。

Image Recognition Technology and Its Application in Nondestructive Detection in Mechanical Parts

Nondestructive Detection is a common method for fault diagnosis, which is pattern recognition in nature. Pattern recognition is considered as image recognition in its narrow sense. Basic concepts of image, image recognition, image recognition process and image acquisition system are introduced. Some common methods for image recognition and their characteristics are compared. The composition of CCD image acquisition system is presented. Image processing and recognition flow are discussed in details by taking the magnetic powder inspection system for crankshaft as an example. Aiming at the situation in which common nondestructive detection system can not meet the detection requirements of crankshaft for high precision, an improved nondestructive detection system is proposed which is applicable to crankshaft. This new detection system is helpful to improve image quality, accuracy and reliability of detection for engine crankshaft.