基于改进U-Net的量子点STM图像分割

为提升量子点图像分割精度,降低特征识别误差,提出一种基于改进U-Net的量子点图像分割方法. 首先,在预处理阶段,设计了以色彩通道为权值的灰度化算法,以提升后续分割效果. 其次,在STM图像分割部分,在原始U-Net结构上引入中间过渡层以均衡网络各层特征. 而后,建立数据集,并通过实验对比不同分割算法的精确度、召回率、F-measure. 最后,将分割算法应用于量子点的特征识别,并测试了不同分割方式对应用的影响. 实验结果显示,改进灰度化方法保留细节信息丰富,明显提升了量子点分割精度;改进U-Net的平均精确率、召回率、F-measure相较原始网络分别提升了13.83%、2.16%、8.13%. 同时,实验数据表明由于分割精度的提升,量子点数量、纵横比等特征参数的识别更加精确.

First-principles Study on Geometric and Electronic Structures of Si（111）-√7× √3-In Surface Reconstruction

In order to determine the structures of Si（111）-√7 √3-In surfaces and to understand their electronic properties, we construct six models of both hexagonal and rectangular types and perform first-principles calculations. Their scanning tunneling microscopic images and work functions are simulated and compared with experimental results. In this way, the hex-H3＇ and rect-T1 models are identified as the experimental configurations for the hexagonal and rectangular types, respectively. The structural evolution mechanism of the In/Si（lll） surface with indium coverage around 1.0 monolayer is discussed. The 4×1 and -√7× √3 phases are suggested to have two different types of evolution mechanisms, consistent with experimental results.

一种新型的STM探针

以普通的石英光纤为材料 ,用熔拉腐蚀复合的方法制备出 nm量级的光纤探针 ,而后在针尖表面镀上数十 nm厚的金属膜 ,达到导电性 ,使其能传导隧道电流 ,从而研制出一种新型的扫描隧道显微镜(STM)光纤隧道探针 ,在 STM上取得了比较理想的实验结果。本文将光纤隧道探针与金属隧道探针作了比较 。