High-throughput pixel-super-resolved coded ptychographic microscopy with a color image sensor

Lensless on-chip microscopy imaging draws widespread attention owing to its evident advantages in simple optical structure, aberration-free imaging, wide field-of-view, and low-cost hardware budget, providing a fertile opportunity for disruptive reductions in cost and revolutionary improvements in portability for biomedical imaging applications. Here, we report a high-throughput pixel-super-resolved coded ptychographic microscopy implemented using a color image sensor. However, the color filtering array (CFA) introduces inherent modulation in the diffraction patterns acquired under monochromatic illumination, leading to spectral crosstalk in the data processing for lensless on-chip imaging. Without manually removing the CFA layer from the sensor chip, we propose a gradient descent optimization method to numerically estimate the transmittance of the CFA layer within the iterative ptychographic phase retrieval process. Compared to schemes that merely use the single-channel data from Bayer-masked diffraction patterns, our proposed method fully utilizes all three-channel data, avoiding the loss of useful and critical pixel information. Experimental results show that a color sensor can achieve high-resolution and high-throughput bio-imaging. This demonstrates the potential for applications involving color sensors in lensless on-chip imaging. Additionally, we developed a cost-effective, compact, and field-portable on-chip microscopy prototype based on the Raspberry Pi single-board computer. We also open-sourced the related control code and reconstruction code for use by interested researchers.