此页面上的内容需要较新版本的 Adobe Flash Player。

获取 Adobe Flash Player

Gray correction method of X-ray fusion image

 

YIN Xiao-gang1, CHEN Ping1,2, PAN Jin-xiao1

 

(1. Shanxi Key Laboratory of Signal Capturing & Processing, North University of China, Taiyuan 030051, China; 2. Key Laboratory of Molecular Imaging, Institute of Automation, Chinese Academy of Sciences,  Beijing 100190, China)

 

Abstract: The conventional X-ray gray weighted image fusion method based on variable energy cannot characterize the physical properties of complicated objects correctly, therefore, the gray correction method of X-ray fusion image based on neural network is proposed. The conventional method acquires 12 bit images on variable energy, and then fuses the images in a traditional way. While the new method takes the fusion image as the input of neural network simulation system and takes the acquired 16 bit image as the output of neural network. The X-ray image physical characteristic model based on neural network is obtained through training. And then it takes steel ladder block as the test object to verify the feasibility of the model. In the end, the gray curve of output image is compared with the gray curve of 16 bit real image. The experiment results show that this method can fit the nonlinear relationship between the fusion image and the real image, and also can expand the scope of application of low dynamic image acquisition equipment.

 

Key words: variable energy; neural network; physical characteristic model; complex structure

 

CLD number: TP391 Document code: A

 

Article ID: 1674-8042(2014)04-0034-06  doi: 10.3969/j.issn.1674-8042.2014.04.007

 

References

 

[1] YANG Pei, DONG Qiu-ying, YANG Min, et al. Bi-energy DR image fusion based on wavelet transform. Nondestructive Testing, 2008, 30(7): 430-433.
[2] YANG Ying, MOU Xuan-qin, ZHANG Min, et al. A technique for extending dynamic range of medical x-ray image. Journal of Xi’an Jiaotong University, 2008, 42(12): 1468-1471.
[3] YANG Ying, MOU Xuan-qin, LUO Tao, et al. Reconstr-uction of X-ray image with super dynamic range. Acta Photonica Sinica, 2009, 38(9): 2435-2438.
[4] CHEN Ping, HAN Yan, PAN Jin-xiao. Research on X-ray multi-spectrum imaging based on variable energy. Spectroscopy and Spectral Analysis, 2013, 32(5): 1383-1387
[5] ZHANG Jian, WANG Yue-wu. The application of neural network model for X-ray image fusion. Computerized Tomography Theory and Applications, 2011, 20(2): 235-243.
[6] SHEN Hui, LIU Zhi-gui, LIU Su-ping. A research on the ray detection based on neural networks. Computer Engineering & Science, 2008, 30(4): 57-60.
[7] CHEN Wen-qing, XU Tang. Application of neural network in the curve fitting of all optical switching of amorphous copolymer containing azobenzence groups. Computers and Applied Chemistry, 2007, 24(5): 1386-1388.
[8] WEI Jiao-tong, CHEN Ping, PAN Jin-xiao. Gradient-energy digital radiography image fusion based on principle component analysis. Chinese Journal of Stereology and Image Analysis, 2013, 18(2): 103-108.

 

 

X射线融合图像的灰度修正方法

 

阴晓刚1, 陈平1,2, 潘晋孝1

 

(1. 中北大学 信息探测与处理山西省重点实验室, 山西 太原 030051;2. 中国科学院自动化研究所 中国科学院分子影像重点实验室, 北京 100190)

摘要:对厚度差异大的结构件, 常规变能量X射线图像融合方法不能正确地表征灰度与物理信息复杂的对应关系。 为此, 本文提出了基于神经网络的X射线融合图像灰度修正方法。 首先, 将常规变能量的融合图像作为神经网络的输入图像, 将16位高动态的图像作为相应的输出图像, 训练得到X射线成像的物理表征模型。 然后, 利用钢质阶梯块验证方法的正确性与可行性, 并将输出结果与16位真实图像进行比较。 实验结果表明, 该方法很好地拟合了融合图像与真实图像灰度之间的非线性函数关系, 扩展了低动态成像采集设备的使用范围。

 

关键词:变能量; 神经网络; 物理模型; 复杂结构

 

引用格式:YIN Xiao-gang, CHEN Ping, PAN Jin-xiao. Gray correction method of X-ray fusion image. Journal of Measurement Science and Instrumentation, 2014, 5(4): 34-39. [doi: 10.3969/j.issn.1674-8042.2014.04.007]


[full text view]