论文部分内容阅读
介绍了一种焊缝X射线图像缺陷深度信息实时自动检测方法。由于要实时和自动 ,则只能用经影像增强器转换的实时X光成像系统 ,而不能用照片方式。又由于现场条件的限制 ,不可能用多幅图的投影合成法。本文介绍了一种仅用三幅图来获取深度信息的方法。区别于照片成像 ,光源已经不能看成平行光 ,文中分析了点光源成像的原理和方法。二维缺陷的准确检测是计算深度信息的基础 ,而准确的边缘检测又是检测二维缺陷的基础。由Marr开创 ,经Canny、Witken和Mallat等人发展起来的多尺度边缘检测被认为是理论上最完善的一种方法 ,但他们没有解决尺度自适应问题。本文介绍了一种基于测度的尺度自适应边缘检测方法 ,获得很好效果。
A method of real-time automatic detection of defect depth information of X-ray image defects is introduced. Due to the need for real-time and automatic, only the real-time X-ray imaging system converted by the image intensifier can be used, not the photo mode. Due to the restriction of the scene conditions, it is impossible to use the projection and synthesis method of multiple images. This article describes a method that uses only three images to get depth information. Different from photo imaging, the light source can not be regarded as parallel light, the paper analyzes the principle and method of point light source imaging. Accurate detection of two-dimensional defects is the basis for calculating depth information, and accurate edge detection is the basis for detecting two-dimensional defects. The multi-scale edge detection, pioneered by Marr and developed by Canny, Witken and Mallat et al., Was considered the most theoretically perfect method, but they did not address the issue of scale adaptation. This paper introduces a measure-based scale-adaptive edge detection method that works well.