Fast autofocusing based on pixel difference with the Tanimoto coefficient between images

Real-time detection and auto-focusing of beam profiles from silicon photonics gratings using the YOLO model

When observing chip-to-free-space light beams from silicon photonics (SiPh) to free space, manual adjustment of camera lens is often required to obtain a focused image of the light beams. In this Letter, we demonstrated an auto-focusing system based on the you-only-look-once (YOLO) model. The trained YOLO model exhibits high classification accuracy of 99.7% and high confidence level of >0.95 when detecting light beams from SiPh gratings. A video demonstration of real-time light beam detection, real-time computation of beam widths, and auto-focusing of light beams is also included.

High-robustness autofocusing method in the microscope with laser-based arrayed spots

Accurate and rapid autofocus technology plays a crucial role in various fields, including automatic optical inspection technology, bio-chips scanning, and semiconductor manufacturing. The current photoelectric autofocus methods have limitations because of detecting the focal plane solely at the center of the microscope field of view. In the application of Stereo-seq the risk of autofocus errors will be increased, which have reduced the robustness of the system, like when the surface of the tested samples are wrinkling and inconsistent thickness, or the detection spot is at the edge of the sample. To enhance the robustness of the autofocus system and mitigate the constraints of the photoelectric autofocus methods, the laser-based arrayed spots photoelectric autofocus method has been proposed. To achieve the uniform light splitting, a 2D-Dammann grating is incorporated into the optical path of the autofocus system, resulting in the formation of an n × n arrayed spots on the surface of the sample. Through experimental verification, it has been demonstrated that this method can achieve the autofocus range of ±100μm and the autofocus accuracy of ±1/4 DOF when applied to a microscope equipped with a 10× objective lens, thereby satisfying the requirements for microscopic focusing. The arrayed light autofocus method devised in this study presents what we believe is a novel research concept for active autofocus detection and holds significant application value.

Fast autofocusing of recorded planes by salient feature region for coherent diffraction imaging

In multi-distance coherent diffraction imaging, the task of distance calculation for multi-diffraction images is cumbersome. The information features are hard-to-extract and the region of interest extraction algorithms are difficult to be adopted. A universal salient feature region selection algorithm by using the area with the highest density of corners is proposed to extract the most representative feature region. In addition, equally spaced recording modes and mismatched diffraction distances will result in system noise and destroy image quality. The polydirectional maximum gradient is offered as a sharpness criterion to weigh a quantitative feature for the final pattern. A fast, sensitive, and high-accuracy autofocusing and sample reconstruction can be achieved using only a small number of images while ensuring that morphological properties and quantification of the reconstructions are not compromised. The proposed method is promising for biological and medical dynamic observations for computational imaging systems.

Simple, non-mechanical and automatic calibration approach for axial-scanning microscopy with an electrically tunable lens

We describe a simple and robust calibration approach for axial-scanning microscopy that realizes axial focus shifts with an electrically tunable lens (ETL). We demonstrate the calibration approach based on a microscope with an ETL placed close to the rear stop of the objective lens. By introducing a target-consisted of repeating lines at one known frequency and placed at a ~45° angle to the imaging path, the calibration method captures multiple images at different ETL currents and calibrates the dependence of the axial focus shift on the ETL current by evaluating the sharpness of the captured images. It calibrates the dependence of the magnification of the microscope on the ETL current by measuring the period of the repeating lines in the captured images. The experimental results show that different from the conventional calibration procedure, the proposed scheme does not involve any mechanical scanning and can simultaneously calibrate the dependence of the axial focus shift and the magnification on the ETL current. This might facilitate imaging studies that require the measurement of fine structures in a 3D volume. We also show the calibration procedure can be used to estimate the radius of a conner-arc sample, fabricated using laser micromachining. We believe that this easy-to-use calibration approach may facilitate use of ETLs for a variety of imaging platforms. It may also provide new insights for the development of novel 3D surface measurement methods. RESEARCH HIGHLIGHTS: The proposed calibration scheme does not involve any mechanical scanning and can simultaneously calibrate the dependence of the axial focus shift and the magnification on the electrically tunable lens (ETL) current. It might facilitate imaging studies that require the measurement of fine structures in a 3D volume, and the use of ETLs for a variety of imaging platforms. It may also provide new insights for the development of novel 3D surface measurement methods.

Depth-of-field expansion method based on multidimensional structure and edge-guided correction

Multi-focus image fusion is a method to extend the depth of field to generate fully focused images. The effective detection of image focusing pixels and the optimization of image regions are the key to it. A method based on multidimensional structure and edge-guided correction (MSEGC) is proposed. The pixel-level focusing evaluation function is redesigned to preserve image details and non-texture regions. Edge-guided decision correction is used to suppress edge artifacts. With public data and semiconductor detection images for verification, the results show that compared with other methods, the objective evaluation is improved by 22-50%, providing better vision.