欢迎访问林业科学,今天是

林业科学 ›› 2018, Vol. 54 ›› Issue (11): 164-171.doi: 10.11707/j.1001-7488.20181123

• 研究简报 • 上一篇    下一篇

木材CT断层成像系统旋转中心校正方法

葛浙东1, 戚玉涵2, 罗瑞1, 陈龙现1, 王艳伟2, 周玉成1   

  1. 1. 山东建筑大学信息与电气工程学院 济南 250101;
    2. 中国林业科学研究院木材工业研究所 北京 100091
  • 收稿日期:2018-04-02 修回日期:2018-05-18 出版日期:2018-11-25 发布日期:2018-12-04
  • 基金资助:
    泰山学者优势特色学科人才团队(2015162);山东建筑大学博士基金"基于X射线的木结构建筑用材无损检测系统研究"(XNBS1622)。

Correction of Rotation Center for the Wood CT Imaging System

Ge Zhedong1, Qi Yuhan2, Luo Rui1, Chen Longxian1, Wang Yanwei2, Zhou Yucheng1   

  1. 1. School of Information and Electrical Engineering, Shandong Jianzhu University Jinan 250101;
    2. Research Institute of Wood Industry, CAF Beijing 100091
  • Received:2018-04-02 Revised:2018-05-18 Online:2018-11-25 Published:2018-12-04

摘要: [目的]构建一套等距扇形X射线束木材CT断层成像系统,提出一种基于正弦图的系统旋转中心校正方法,实现对木材高质量断层成像,为准确获取木材内部结构信息、研究木材结构形态提供技术支持。[方法]首先,分析CT断层成像系统成像原理,验证穿过扫描断层内任意一点的投影地址与旋角度存在一定函数关系,投影地址轨迹在扫描过程中是一条与旋转角度相关的正弦曲线;其次,计算投影地址对称中心与探测器中点的差值,在图像重建过程中,对旋转中心偏移量进行补偿,获得任意一点正确的投影位置;再次,选择直径205 mm、高度400 mm的杉木作为试验对象,分别校正旋转中心偏移量为8、9、10、11、12个探测点,对断层1进行扫描,获得5幅断层图像,并对试验结果进行定量分析;最后,对断层2进行扫描,重建旋转中心校正前后断层图像,并对2幅图像进行评价与分析。[结果]断层1的5幅重建图像均可反映木材内部结构特征,校正数值逐渐接近旋转中心偏移量,峰值信噪比和结构相似度均有所提高,而相对误差均有所减小,图像愈加近似于原始图像,旋转中心校正数值与偏移量相等情况下重建图像效果最好。断层2的2幅重建图像质量差异性较大,校正前图像环形伪影干扰严重,生长轮肉眼难以区分,图像边缘有向周边扩散现象,相近裂纹结构产生假象,而校正后图像无环形伪影干扰,心材生长轮清晰可见,图像边缘无向周边扩散现象,成像质量良好。旋转中心校正后的图像质量远远优于校正前,基于正弦图的系统旋转中心校正方法在木材CT断层成像系统中具有良好的应用效果。[结论]基于正弦图的系统旋转中心校正方法可有效解决木材CT成像系统旋转中心随机偏移问题,抑制环形伪影缺陷,明显提升木材CT断层图像的重建效果。木材CT断层成像系统旋转中心经校正后,重建的木材断层图像可清晰反映裂纹、生长轮、心边材等结构特征,能满足科研、高校、检测检验等机构对木材内部结构的检测与分析要求。

关键词: X射线, CT, 木材, 正弦图, 旋转中心

Abstract: [Objective] A computed tomography(CT) imaging system for wood based on equidistant fan-shaped X-Ray beam was developed and a method of correcting the rotation center based on the sinogram was presented in this paper. Applied the CT imaging system in wood,the reconstructed images of wood could be used to facilitate universities and scientific research institution, and to provide technical support for the structural morphology research.[Method] Firstly, the image reconstruction principle of the CT system was analyzed to verify that there was a functional relationships between the projection address of any point in the scanning fault and the rotation angle. The trajectory of projection addresses was a sinusoidal curve associated with the rotation angles during scanning process.Secondly, the difference between the center of symmetry of the projection address and the midpoint of the detector was calculated. In the process of image reconstruction, the offset of rotation center was compensated to obtain the correct projection position of any point.Thirdly, a Chinese fir with the diameter of 205 mm and the height of 400 mm was selected as the sample, and the offset of rotation center was corrected by 8, 9, 10, 11 and 12 detecting pointsrespectively. Five reconstructed images were obtained by scanning the fault 1 of the sample, and these images were analyzed quantitatively. Finally, the fault 2 of the sample was scanned to reconstruct two images before and after the correction of the rotation center, and the images were evaluated and analyzed.[Result] The 5 reconstructed images of fault 1 could all reflect the internal structural characteristics of wood. As the corrected values getting closer to the offset of rotation center, the PSNR and SSIM values increased, and the RE values were decreased. It showed that the reconstructed images were more similar to the original image. In order to obtain the best reconstructed image, the corrected value of the rotation center must be equal to the offset. The two reconstructed images of fault 2 were quite different. For example, quite a lot of ring artifacts of the reconstructed image before correction brought many difficulties in distinguishing the growth rings. The edge marks diffused to the periphery, and the false similar crack structures appeared in the reconstructed image. In the corrected image, there weren't ring artifacts or peripheral edge marks. Growth rings in the heartwood could be seen clearly, and the reconstructed image was perfect.The experimental result show that the image at the corrected rotation center was much better than the pre-correction image. The method of correcting the rotation center through the sinogram has a good application effect in the wood CT tomography system.[Conclusion] The method for correcting rotation center was proposed in this paper. It effectively solved the problem on random migration of the rotation center, suppressed the ring artifacts, and improved the quality of reconstructed images. After the rotation center is corrected, the reconstructed images can show internal structural characteristics suchascracks, growth rings, heartwood and sapwood much more clearly.The CT imaging system meets the requirements of universities and scientific research institutions for detecting and analyzing the internal structural characteristics of wood.

Key words: X-ray, computed tomography (CT), wood, sinogram, rotation center

中图分类号: