摘要
为了对X射线安检设备的成像质量进行客观评价,根据GB 15208-2005标准中图像质量评价的方法和要求,参考主观评价的过程,提出了图像处理和模式识别的方法。对测试体A采用基于ROI特征提取的处理方法,对测试体B采用基于色彩信息特征提取的处理方法。介绍了两种方法的处理原理、流程和关键步骤,设计了实现上述算法的软件系统。试验结果表明,在图像存在噪声和畸变等情况下,推荐的算法仍然具有较高的正确率,能够遵循标准的要求对图像质量进行客观评价。
For evaluating the image quality of X-ray security inspection system objectively, according to the testing methods and requirements about image quality evaluating of standard GI3 15208--2005, referring to the process of subjective evaluation, the image processing and mode recognition methods were put forward. ROI-based feature extraction method was carried out for testing block A, while color information based feature extraction method was carried out for testing block B. The principles, processes and key steps of the two signal processing methods were introduced, and thus software system was designed and achieved. Testing results showed that the recommended algorithm was high in accuracy even while the images being of noise and distortions, so it could evaluate the image quality objectively according to the standard.
出处
《无损检测》
2009年第6期428-432,共5页
Nondestructive Testing
基金
民航安全技术分析和鉴定实验室开放研究基金资助项目
关键词
X射线安检设备
人眼视觉特性
图像质量
特征提取
X-ray security inspection system
Human visual characteristics
Image quality
Feature extraction
