磨粒图像拼接是铁谱分析技术中遇到的难题之一。为了解决该问题,介绍了图像拼接的原理,分析了磨粒图像拼接的特点。在介绍了各种图像拼接方法优缺点的基础上,分析了适合于磨粒图像的拼接方法,并提出了改进算法,综合考虑了算法的速度和算法效果,给出了实现方法与途径。实验结果表明,其中基于比值的方法程序实现容易,速度快;基于图像相关性的配准方法稳定可靠,加上基于直方图的模版自适应选取,提高了磨粒拼接的自动化程度。
Wear particle image mosaicing is important for ferrography.In order to solve this problem,mosaicing theory and wear particle image characteristic is introduced.Analyze the mosaicing methods and present the improved algorithm based on advantage and disadvantage in image mosaicing methods.Considering the balance between algorithm speed and effect,mosaicing methods and its realizing steps for wear particle images are presented.Experimental results show that,the method based on the ratio between different pixels in images is easy and fast for computer processing,while the method based on image relativity is steady and reliable.Using histogram,template automatic selecting for wear particle image mosaicing is realized.