摘要
针对人眼观察固体火箭发动机ICT图像容易误判及遗漏缺陷的问题,采用ICT图像配准技术,将待检固体火箭发动机的ICT图像与无缺陷标准ICT图像配准对比,以便突出固体火箭发动机装药缺陷。改进了基于归一化互信息量为配准测度的配准方法,在配准测度中添加进了ICT图像像素之间的位置信息,以增强配准算法的精度和鲁棒性。实验结果证明,提出的方法可以精确的完成图像的自动配准并可使缺陷部位灰度对比更加显著,便于捕捉缺陷。
In order to help engineers identify solid rocket motor＇s inner faults when they check solid rocket motor（SRM） ICT images, a method based on the registration theory of ICT images is used. SRM＇s inner faults could be highlighted when their ICT images are compared to the standard ones. The mutual information registration theory is improved by counting in the position information of pixels in ICT images, and accuracy and Robust character are increased. At last, the floating image is compared with the reference image and the faults＇ characters as well as their positions are achieved. The results show that the method can complete the registration of ICT images well and the inner faults are easier to seize after comparison.
出处
《战术导弹技术》
2014年第5期82-86,共5页
Tactical Missile Technology
