基于图像分析的射线照相缺陷识别技术在无损检测领域被越来越广泛的使用。其一般步骤是：首先通过高分辨率摄像机拍摄被测焊缝的底片，然后对采集到的图像进行图像去噪和增强等预处理操作，再利用模式识别等手段识别其中的主要缺陷。可见，图像增强是其中一种非常有用的图像处理工具。在焊缝图像中，图像增强的主要目的是增强焊缝缺陷和灰度背景之间的对比度，以便于后续的分割和模式识别。研究了一种基于局部二值特征焊缝图像增强算法，并应用于焊缝图像的缺陷识别系统中。试验结果表明使经过增强处理后的图像，缺陷特征明显，非常有利于进一步的缺陷提取和识别。

Radiograph testing based weld defect recognition has been widely used in nondestructive testing of the weld defect. The main steps of the recognition are as follows. Firstly, high-resolution video camera is used to capture the tested RT film, following the image denoising and enhancement preprocessing operations, then pattern recognition and other means are used to identify the main defects in the film. Image enhancement is one of the very useful image processing tools. The purpose of image enhancement is to strengthen the contrast between the defects and background, in order to facilitate the subsequent segmentation and defect pattern recognition. In this paper, we propose a weld image enhancement algorithm based on the Local Binary Pattern(LBP), and apply it into the weld defect image recognition system. Experiment result shows that, after such image enhancement, the defects in the image become clearer and could be recognized better than without enhancement.