X光图像自动增强拼接技术的研究和实现

Research and Implementation of Automatic X-ray Image Enhancement and Stitching Technology

目的:设计了一种X光图像的自动增强拼接技术。方法:首先,针对X光影像灰度差异目测困难,单一直方图均衡化增强后可能导致的灰度级减少,某些细节消失,提出一种改进的基于HE的图像增强方法,利用三角函数加权因子对图像进行平滑增强融合;然后,为了解决由于X光图片视野范围的局限性,采用基于相位相关法对图像进行全景拼接重建,并分析了参考图像与待配准图像间存在重叠区域的面积大小与拼接成败的关系;最后,为了提高图像的拼接质量,采用多分辨率算法进行图像融合。结果:成功实现了图像自动增强与拼接。结论:实验证明,对于X光图像存在的灰度差异微小,细节不明显的情况,经过改进的图像增强之后,效果明显,且方法运算简单,能快速有效地对X光图像进行自动增强及拼接及融合,鲁棒性好。当参考图像与待拼接图像重叠区域减小到参考图像2%以下时,才出现误拼接,相同重叠区域面积内,图像信息携带量越多,可得到更好的匹配效果。

Objective： To design an automatic enhancement and stitching technology of X-ray images. Methods： Firstly, present an improved image enhancing method based on HE, which registry and make images smoothly by weight of trigonometric function. It resolves these problems which includes X-ray image with tiny differences in gray, reducing gray levels and missing details after image enhancing only by HE. Secondly, use a method based on phase correlation technolo- gy to stitch images to enlarge the eye shot of X-ray image, and analyzing the relationship between the size of the images＇ overlapped area and whether the stitching is successful. Lastly, registry images seamlessly with multi-resolution algorithm. Results： Achieve the images automatic enhancement and stitching. Conclusions： The experiment has proved that the tech- nology can improve the case that X-ray image of the existence of small differences in gray and inconspicuous details. What＇s more, it can automatic enhancement and stitching X-ray images fast and effective. And the method has strong robustness. It＇s not successful until the size of the images＇ overlapped area lower than 2 % of image. What＇s more, the information of image takes more in the same size of the overlapped area, the better stitching effect.