计算机视觉
像素
人工智能
高光谱成像
旋转(数学)
翻译(生物学)
计算机科学
材料科学
霍夫变换
图像分辨率
模式识别(心理学)
基础(线性代数)
帧(网络)
图像处理
图像(数学)
数学
几何学
生物
信使核糖核酸
基因
电信
生物化学
作者
Kirby Schmidt,Anthony Trofe,Tetyana Ignatova
标识
DOI:10.1149/2162-8777/aca99c
摘要
It is common to characterize a single sample using multiple different microscopy methods that have variable scales, rotation and translation. Registering hyperspectral images of 2D materials is particularly difficult due to the lack of keypoints on un-patterned substrates. The strain/doping maps can assist by creating these keypoints. Registration of these images allow for multimodal analysis from these various instruments by aligning multiple images into a single coordinate space. This is done by Hough transformations and arbitrary resolution definitions to generate a new coordinate frame where spatial information may be preserved and correlated on a pixel by pixel basis.
科研通智能强力驱动
Strongly Powered by AbleSci AI