论文部分内容阅读
针对基于扫描电子显微镜(SEM)的纳米操作系统中,SEM图像噪声大,难以实现小目标物体(纳米线)的识别等问题,利用核密度估计来识别纳米线,以实现其精确定位.首先,利用核密度估计实现阈值分割,以便实现纳米线提取;然后,利用图像处理中的细化算法,获取纳米线的中线方程;最后,利用纳米线的中线方程和纳米线图像,用基于梯度方向高斯曲线拟合方法实现对纳米线的精确定位.实验数据表明,核密度估计可以实现对纳米线的准确识别,在纳米线的定位方面具有很好的准确性和可靠性.
For the nanometer operating system based on the scanning electron microscope (SEM), the SEM image has a large noise and it is difficult to identify the small target (nanowire). The nuclear density estimation is used to identify the nanowire to achieve its precise positioning.Firstly, The threshold segmentation is implemented by using kernel density estimation in order to realize the extraction of nanowires. Secondly, the midline equation of nanowires is obtained by using the thinning algorithm in image processing. Finally, by using the midline equations of nanowires and nanowire images, The curve fitting method can realize the precise positioning of the nanowires.The experimental data show that the nuclear density estimation can accurately identify the nanowires and has good accuracy and reliability in the nanowire positioning.