Adaptive AFM imaging based on object detection using compressive sensing.

Area of Science:

• Nanotechnology
• Materials Science
• Surface Science

Background:

• Atomic Force Microscopy (AFM) provides high-precision nanoscale surface morphology measurements.
• Standard AFM scanning is time-consuming, limiting high-resolution imaging speed.
• Traditional Compressive Sensing (CS) in AFM struggles with uniform image quality across different sample areas.

Purpose of the Study:

• To develop a novel adaptive compressive sensing (CS) imaging scheme for faster and higher-quality Atomic Force Microscopy (AFM).
• To address the limitations of traditional CS-AFM in balancing local image quality, particularly in object areas at low sampling rates.

Main Methods:

• Implemented a fast scanning protocol followed by bicubic interpolation to generate an initial high-resolution image.
• Utilized an advanced detection algorithm for accurate object localization.
• Performed supplementary adaptive sampling focused on detected objects.
• Constructed a measurement matrix for the sampled points.
• Reconstructed the complete AFM image using Total Variation Minimization by Augmented Lagrangian and Alternating Direction Algorithm (TVAL3).

Main Results:

• The adaptive CS-AFM scheme demonstrated improved image quality, assessed by PSNR and SSIM metrics and visual inspection.
• Achieved high automation, reduced imaging time, and superior image quality compared to non-adaptive methods.
• Successfully balanced image quality across different local areas, enhancing detail in object regions.

Conclusions:

• The proposed adaptive CS-AFM scheme offers a significant advancement in fast, high-resolution nanoscale imaging.
• This method effectively overcomes the trade-offs between speed and quality in traditional CS-AFM.
• The adaptive sampling strategy ensures detailed imaging of important features without compromising overall efficiency.