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 view of the human eye to observe the easy misjudgment and missing defects of solid rocket motor ICT images, the ICT image registration technology is adopted to compare the ICT images of the solid rocket motor to be tested with the standard ICT image without defects in order to highlight the solid rocket motor Drug defects. In order to enhance the accuracy and robustness of the registration algorithm, the registration method based on normalized mutual information is improved. The registration information is added to the location information of ICT image pixels. The experimental results show that the proposed method can accurately complete the automatic registration of the image and make the contrast of the gray level of the defect more significant, so as to capture the defect.