AFM Sample Position Identification Using Vision and Prediction Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The usability of atomic force microscopes is degraded by manual setting of measurement positions, and there is a need for an automated method to accurately identify sample positions, especially during calibration steps.
Innovation Solution
A machine-learned model and apparatus that uses a vision unit and a prediction model, such as an artificial neural network, to identify sample positions based on vision images, allowing for accurate and automatic setting of sample positions without user intervention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If manual setting of measurement position is used, then the atomic force microscope can operate with simple equipment, but the usability is degraded and operation becomes time-consuming
Solution Approach 1:
The system performs automatic self-positioning by capturing images, identifying sample regions through machine learning, and autonomously adjusting the stage to center the sample. This eliminates the need for manual position setting by the user, directly improving usability and reducing time loss.
Solution Approach 2:
The patent replaces manual mechanical positioning with an automated vision-based system. A camera captures images, a machine learning model identifies sample positions, and the system automatically adjusts the stage, substituting human manual operation with an automated optical and computational system.
2Ease of operation
If automatic identification of sample position is implemented, then usability is improved, but the device complexity increases due to additional vision units and prediction models
Solution Approach 1:
The vision unit serves multiple functions: capturing sample images for position identification, providing visual feedback to users, and enabling automated positioning. This multi-functionality reduces the need for separate dedicated components, thereby limiting the increase in device complexity.
Solution Approach 2:
The machine learning prediction model acts as an intermediary between the raw image data and the control system. It processes images to identify sample regions and provides position information to the stage controller, simplifying the overall system architecture by decoupling image processing from control logic.
3Measurement precision
If machine learning model is used for sample position identification, then identification accuracy is improved, but computational resources and processing time are increased
Solution Approach 1:
The machine learning model is trained in advance on a dataset of sample images to learn sample position patterns. This preliminary training allows the model to quickly and accurately identify sample positions during actual operation without requiring complex real-time computations, thereby reducing power consumption during use.
Solution Approach 2:
The system processes only the most relevant features of the image for position identification rather than analyzing the entire image in detail. This partial processing approach maintains high identification accuracy while reducing computational resource requirements.
Data Source
AI summary
An apparatus and a method for identifying a sample position in an atomic force microscope according to an exemplary embodiment of the present disclosure are provided. The method for identifying a sample position in an atomic force microscope includes receiving a vision image including a subject sample through a vision unit; determining a subject sample region in the vision image using a prediction model which is configured to output the subject sample region by receiving the vision image as an input; and determining a position of the subject sample based on the subject sample region.