This PDF file contains the front matter associated with SPIE Proceedings Volume 12942, including the Title Page, Copyright information, Table of Contents, and Committee Page.

The phase recovery algorithm based on the transport of intensity equation uses the fast Fourier solution to calculate the phase from the acquired intensity, but the solution accuracy is not high, and there will be instability caused by zero points and minimum points. Aiming at this problem, An improved fast Fourier solution based on the intensity transfer equation is proposed. By finding a suitable constant value to replace the focused intensity value in the traditional formula, the initial guess solution of the phase is solved; the initial phase and the focused intensity form a new complex amplitude, and then a new intensity differential is obtained in the form of angular spectrum propagation, and then the new The intensity differential of is substituted into the phase solution formula to obtain a new phase, so as to iteratively optimize the phase; when the iteration converges, the exact solution of the phase can be obtained. This solution can bypass the instability caused by the zero point and the minimum value point and has the advantage of high precision. Keywords: Transport of intensity equation, Intensity differential, Iterative optimization, Angular spectrum propagation, Fast Fourier solution, phase recovery.

The binary defocusing technique is sensitive to the defocusing degree. The defocusing projection mechanism will introduce high-frequency harmonics at the inappropriate defocused level, leading to limitations in measurement accuracy and depth range. In this paper, a binary-focusing projection technique combining generative adversarial networks is proposed. First, the focusing binary patterns based on error diffusion are projected on the measured surface, and then the captured fringe patterns are input to generative adversarial networks, which achieves sinusoidal correction and optimization for both the focused region and the low-quality defocused region due to its strong image translation ability. Finally, 3D measurement is realized by a phase-shifting algorithm. Compared with the traditional binary defocusing technique, the proposed method is not limited by the defocusing degree and maintains the advantages of high-speed projection, so it can achieve a larger measured depth range and improve measurement accuracy. Simulation and experiments verify the performance of the proposed method.

Based on the extremely narrow half-height width (FWHM) characteristics of ultrashort pulsed lasers, a DIC imaging technique that can realize picosecond time resolution is proposed by combining the traditional digital image correlation (DIC) technique with it. This method can solve the problems of blurring and dragging of scattering images caused by the sudden crack initiation and fast crack expansion in the fracture process of brittle materials. The crack opening displacement (COD) and the full-field displacement of the sample at the instant of cracking of brittle materials can also be obtained. In addition, the error in spatial resolution of the traditional DIC technique using a continuous light source can be greatly reduced by using this method. In this paper, this method is used to record the tip displacement field of a tuff sample containing a type I prefabricated crack under semicircular disk three-point bending (SCB) experimental conditions. The experimental results show that the recording of crack extension behavior on the order of picoseconds can be achieved using this method, and the key parameters of fracture of brittle materials are calculated more accurately.

Phase retrieval algorithms, such as the Wirtinger Flow (WF) algorithm, are widely used in various fields. As a nonconvex optimization algorithm for phase retrieval, WF is commonly employed in the reconstruction of holograms in holographic image projection. These types of algorithms typically involve two stages: an initialization stage and an iterative optimization stage. In the initialization stage, an initial value is provided, and a spectral method is used to calculate an approximate solution as the initial guess. The iterative optimization stage then utilizes the Wirtinger gradient to iteratively compute and converge the initial guess to a nearby real solution, thereby obtaining the global optimal solution. However, due to the random nature of the initial values, the computed results often exhibit significant instability. To address this issue, this paper proposes an approach based on a quadratic distribution for improving the stability of the results. In the initialization stage, the initial value is set as the quadratic distribution initial value. Then, the spectral method is applied again to calculate the initial guess. Since the quadratic distribution initial value is artificially assigned, it enhances the stability of the computed results. To validate this method, the paper applies the quadratic distribution initial value to both the initialization stage of the WF algorithm and the Truncated Amplitude Flow (TAF) algorithm. A comparison is made between the results obtained using random initial value and those obtained using the quadratic distribution initial values. The results demonstrate that compared to random initial values, the quadratic distribution initial values can achieve faster and equally accurate computation results with higher stability. Finally, this method is applied to simulation experiments of in-line digital holography, and the reconstruction results from the experiments further confirm the effectiveness of our approach.

For free space optical transmission, vortex beam can greatly improve the channel capacity, but it is easy to be affected by atmospheric turbulence. To solve this problem, in this paper, a radially aligned Gaussian beam array is loaded with a discrete vortex phase-coherent synthetic vortex beam. Based on the multi-phase screen numerical simulation method, the transmission of the synthetic vortex beam in Von Karman spectrum atmospheric turbulence is simulated, and the intensity distribution, drift and flicker characteristics of the beams under different turbulence intensity are studied. The effect of topological charge on light intensity flicker and beam drift is also discussed. The simulation results show that when the coherent vortex beam is transmitted in atmospheric turbulence, the turbulence makes the intensity distribution at the receiving end disordered and the phase distribution distorted. With the increase of turbulence intensity and transmission distance, the scintillation index and drift mean square error will increase, but when increasing to a certain extent, the scintillation index will tend to be flat. Under the same transmission conditions, the more topological charge of the beam, the better the transmission quality of the coherent synthetic vortex beam.

In recent years, the lithium battery industry has been developing rapidly, and in the process of its large-scale industrialized production, the automatic defect detection technology based on machine vision has extremely important research value. Because of the complexity of the lithium battery production environment, the defect morphology is variable, the current research results for lithium battery pole piece defect detection is relatively small. In order to meet the needs of lithium battery pole piece defect detection speed and accuracy, to solve the problems of complex background noise, defects and low contrast in the pole piece image, this paper proposes a lithium battery pole piece defect detection algorithm based on machine vision technology, firstly, adopt the topological mapping based on the weighted average neighborhood closure curve filtering template for the image noise reduction processing, and then use the wavelet transform based on the multiscale detail enhancement method for image enhancement processing;; subsequently, adopt the multi-scale detail enhancement method based on wavelet transform for image enhancement processing; and subsequently, use the topological mapping based on the weighted average neighborhood closure curve for image enhancement processing. Then, in order to solve the problem of uneven illumination and more speckle impurities in the polar film image, the area growth method is used and combined with differential geometry tools to extract the defect contour of the area to be tested; finally, the concept of Earth Move Distance (EMD) is introduced, which is used to compute the similarity between the obtained contour and various types of defect templates contours to realize the classification of defects. Experiments have shown that the algorithm in this paper improves the speed and accuracy of defect detection on the surface of the pole piece, retains the details of the defect edges, detects small defects with low contrast, and extracts the complete defect contour, which better meets the actual needs of industrial production.

Monocular imaging is constrained by limitations in the detection angle, making it susceptible to visual deceptions and making it difficult to obtain accurate shape and structural information of three-dimensional objects. The polarimetric characteristics of scattered light from objects contain information about surface roughness, texture, and structural differences. Therefore, introducing polarization measurements into monocular imaging systems holds significant potential. In this paper, based on polarized 3D imaging theory, the acquisition of surface normal information of objects is achieved by establishing the Stokes vector equation and relating it to Fresnel reflection and the Malus law. Rendering of normals and object surface directions is performed in the 3DsMax software. Ultimately, a monocular visual polarization imaging method is employed to correct the visual deception effect of objects with deceptive features. The results demonstrate that this method exhibits a certain recognition ability for three-dimensional objects composed of multiple planes with deceptive viewing angles.

Light field 3D display is a glasses-free 3D display that samples and reconstructs the light field guided by ray optics. To increase the density of the discrete viewpoints, parallax images in more than tens or even hundreds of directions should be sampled, which adds up to computation and slowdowns the generation speed of the 3D image source. In this paper, a fast generation method for 3D image source based on instancing camera rendering is proposed, which greatly optimizes the speed of the parallax image rendering process in light field 3D display. Experimental results reveal that the proposed method can generate the 4K and 8K light field 3D image sources at over 60 frames per second (fps) with viewpoints less than 160 and 80, respectively, performing at least 106% and 94% faster than the conventional non-instancing method.

The imaging depth of conventional Optical Coherence Tomography (OCT) is limited by high scattering of biological tissues, while the signal intensity of deep tissue imaged by Photoacoustic Microscopy (PAM) is also affected by the weak light excitation of biological tissues. In this paper, glycerol solution was used as the optical clearing agent (OCA) to enhance tissue transparency and reduce light attenuation during deep tissue imaging. We performed optical clearing treatment on the anterior and posterior segments of rabbit eyes by topically applying glycerol to the conjunctival opening and through posterior injection, respectively. Then the anterior and posterior segments of rabbit eyes were imaged using the PAM and OCT systems. The results demonstrate that the optical transparency alteration of the anterior and posterior segments of rabbit eyes changes the tissue refractive index, increases the signal intensity of OCT and PAM, and enhances the imaging depth of both OCT and PAM. Consequently, the optical clearing agent provides a powerful tool for ophthalmic research and early diagnosis of ocular diseases, and also expands the imaging applications of OCT and PAM.

Spark energy is one of the most important indicators for evaluating the performance of electric ignition systems. The development of electric ignition systems requires accurate measurement of spark energy to optimize system design parameters.The traditional oscilloscope test method calculates the spark energy by measuring the discharge electric energy. There is a serious energy conversion error, and the accurate measurement of spark energy cannot be realized. In this paper, based on radiation energy detection, a new method for direct measuring spark energy is reported. The multiband photodetector is used to conduct spatial sampling and spectral integration of the spark radiation energy. Then, using the high-speed response capability of the photodetector, high-precision measurement of spark energy is achieved by combining the time domain waveform of the spark pulse with the time integration of the spark radiation power. The experimental system can sample the spark radiant energy in 200nm~12,000nm spectral range by using 12 photodetectors, which is divided into four wavebands, and realize the direct test output of spark energy. The energy testing results show that the precision and stability of spark energy measurement are better than 5%. On the one hand, the method utilizes photodetectors to detect the radiation energy produced by electrical sparks and directly obtain the spark energy without requiring conversion between different forms of energy. Therefore, this approach offers higher measurement accuracy. On the other hand, the method takes advantage of the natural electromagnetic interference immunity of optical measurement techniques, which can effectively address the issue of strong electromagnetic interference caused by the electrical ignition system in oscilloscope methods. This can prevent distorted test results and ensure the ability to complete normal tests. Further studies show that the method can be used for accurate measurement of spark energy.

As we all know, the traditional compressed light field 3D display technology has the problems of limited 3D depth of field and low display brightness. In this paper, a hybrid compressed light field device based on polarization multiplexing is proposed, which combines multiplicative and superimposed compressed light field 3D display to improve the light intensity perceived by human eyes and enlarge the depth of field. In addition, when using high-brightness mini-leds, noise can appear at the edges of the reconstructed image. This is because non-negative tensor matrix (NTF) algorithm adopts hierarchical iteration, which is easy to fall into the local optimal solution, resulting in poor optimization effect of the edge part and noise. Then we introduce the stochastic gradient descent (SGD) algorithm which can better improve the problem of edge noise because all spatial light modulator pixel values are updated at the same time in the iteration process. In terms of perception indicators, NTF uses the mean square error coefficient, which cannot account for many nuances of human perception, resulting in iterative results that sometimes do not conform to the subjective perception of human eyes. In contrast, the loss function of SGD can be self-defined. This paper introduces the Learned Perceptual Image Patch Similarity, which is more in line with human perception. Through simulation and experiments, we verify the advantages of the proposed device and the effectiveness of the corresponding optimization algorithm.

Compressive light field (CLF) is a promising light field display technology, and the traditional multiplicative CLF limits the number of layers due to the low transmittance of liquid crystals, which results in a small depth of field. Therefore, this paper proposes a three-dimensional display structure with a hybrid CLF. This structure utilizes a semi-transparent and semi-reflective mirror to superimpose two sets of multiplicative CLFs, each of which consists of two identical liquid crystal displays and a uniform backlight. The hybrid CLF has a greater depth of field and higher brightness, further improving image quality. Due to the properties of the hybrid CLF structure and the non-negative tensor (NTF) decomposition algorithm, the reconstructed image can suffer from layered image crosstalk, which leads to image quality degradation. We propose a method to reduce the hybrid CLF layered image crosstalk, and we validate the proposed method through computer simulations and optical experiments.

With the increasing precision and complexity of high-density interconnect integrated circuit (IC) substrates, automated visual inspection encounters significant challenges in accurately detecting etching defects on metallographic substrate images. Factors such as grayscale variations, noise interference, and rich textures further complicate the process. To address this issue, a novel detection method based on differential geometry theory is proposed, encompassing defect detection between circuits and on circuits. Firstly, the variational Chan-Vese model and morphological closing operation are employed to obtain highly accurate substrate segmentation images. For defect detection between substrate circuits, contour regions between circuits are extracted by differencing the original image with the segmented image. Next, a lightweight compressed MobileNet (CMNet) network is constructed using depth-weighted compression to rapidly identify defect regions between circuits. For defects on substrate circuits, the contour of the segmented image is utilized to determine candidate regions of etching defects by evaluating abrupt changes in angles between adjacent contour points. Subsequently, the proposed discrete curvature calculation method based on the Frenet frame of differential geometry theory is employed to detect and measure defect candidates on the circuits. Experimental results demonstrate the effectiveness of the proposed method in detecting etching defects, outperforming other advanced techniques in screening and identifying defect regions.

In recent years, significant progress has been made in deep learning-based image deblurring. These approaches utilize deep neural networks to learn the map between blurry and clear images or jointly learn the blurry kernel and clear image. They have demonstrated effectiveness in enhancing image quality, preserving details, and handling various types and degrees of blur. The objective of this study is to develop a defocus enhancement technique for real-world scenarios using score-based generative models. Stochastic differential equations (SDE) are employed to gradually introduce noise, thereby smoothing the data distribution towards a known prior distribution. The Score-Matching Langevin Dynamics (SMLD) model estimates the score for each noise scale, while Diffusion Models (DDPM) train the target model for score computation. This process constructs a score-based model capable of reversing the SDE over time. A predictor-corrector framework corrects the evolution of the reverse-time SDE, and the prior distribution is transformed back to the data distribution by removing the noise. By leveraging score-based generative models, accurate score estimation and sample generation are achieved using neural networks and numerical SDE solvers. This technique effectively restores clarity and details in defocused images, thereby enhancing overall image quality.

Quantitative phase imaging and measurement of surface topography and fluid dynamics for objects, especially for moving objects, is critical in various fields. Phase-shifting digital holography, as a highly accurate phase measurement technology applied for moving objects, is limited by some aspects, such as dynamic phase measurement, accuracy of phase shift and temporal phase sensitivity. In this study, we proposed a two-stage neural network (VY-Net) for one shot phase recovery. This Y-Net generates two holograms with specific phase shifts from a single-frame phase shifted hologram, then V-Net recovering the phase with the three holograms input. Simulation results prove that the proposed method can provide an alternative approach for systems of phase-shifting digital holography based on common-path configuration to realize rapid phase-shifted holograms acquisition and accurate phase measurement.

Optical bottle beams exhibit a periodic cosine distribution along the axial direction. It has found extensive applications in areas such as optical trapping, optical tweezers, and guidance of biological cells. However, the current methods employed for generating such cosine light fields lack flexibility in controlling the period and phase, thus failing to achieve desired customization. This paper presents a novel approach for generating customized cosine light fields based on dual diffractionfree beams. The first step involves defining the desired cosine distribution light field. Subsequently, using the annular aperture method, multiple diffraction-free Bessel beams with different axial wave vectors are generated, and their superposition approximates the desired cosine light field. By modifying the phase and period of the desired cosine light field, adjustments can be made to the radius and width of the annular aperture, thus easily obtaining a customized cosine light field. The feasibility of the method was validated through rigorous mathematical analysis, and its effectiveness is demonstrated through experimental verification. This method is expected to propel further applications of optical bottle beams in areas including atomic trapping, optical modulation, and guidance of biological cells.