Fractional-domain modulated microscopic 3D reconstruction method

Key Points

• The study aims to develop a cost-effective and high-precision 3D reconstruction system using multiscale adaptive fractional methods.
• Developed a multiscale adaptive fractional model for depth (MAFD) system combining hardware and software.
• Implemented an adaptive sparse sampling algorithm to manage memory overflow from high-resolution images.
• Created a sequence matching algorithm guided by full-focused images to address camera jitter issues.
• Introduced an adaptive fractional Fourier transform module to enhance sensitivity to noise and detail in images.
• The proposed system achieved greater reconstruction accuracy compared to existing methods.
• Demonstrated significant efficiency and cost advantages over laser confocal and structured light 3D reconstruction methods.
• Showed robustness to noise and improved depth estimation in real micro scenes.