高速粒子场的全息再现图像的自动分割方法

针对再现图像的处理,通过将图像细分为多个较小区域,在每个小区域同时采用多种阈值计算方法,并引入错误校正机制,获得了能稳定地分割图像的自适应阈值.在此基础上,再用种子生长方法进一步界定粒子的边界,能获得更准确的分割结果,且其过程可以自动完成.该方法已成功用于动态实验图像的处理,获得了很好的结果,静态粒子靶实验表明该方法处理得到的相对误差小于15%.

An Automatic Segmenting Method for the Reconstructed Image of High Speed Particle Field

In the measurement for high speed particle field ejected from material surface under explosion,the size,shape,position,and velocity of the particles could be obtained by pulsed holography technique.The reconstructed images were properly segmented by using several threshold selection methods together and adding an error correction in every part,after divided the image into several small parts.On the base the particles edge ware determined by region growing.And all of the above work could be accomplished automatic in computer.The famous results have been got when used in dynamic test.The experiments on static particle target indicates the error are less than 15%.