Computational simulation of multi-wavelength light-field thermometry based on a chromatic meta-lens.

This Letter proposes a light-field meta-lens multi-wavelength thermometry (MMT) system that is capable of modulating a full-spectrum incident radiation into four separate wavelength beams. The chromatic meta-lens is designed using finite-difference time-domain (FDTD) software to function as a filter, ensuring its ability to separate four wavelengths. The chromatic meta-lens is positioned on the back focus plane of the main lens to replace the microlens used in traditional light-field systems and simplify the overall system. After detecting the acquired wavelengths and intensities of the image on photodiodes, a raw multispectral image can be decoupled and processed using the Chameleon swarm algorithm (CSA). Four full-spectrum incident radiations corresponding to four temperature characteristic curves are detected. The high accuracy of the reverse temperature calculation enables the measurement of surface high-temperature distribution with precision.

Chameleon swarm algorithm for data processing of a light-field multi-wavelength pyrometer.

It is recognized that unknown emissivity and ill-posed radiation equations present significant challenges to light-field multi-wavelength pyrometry. Furthermore, emissivity range and choice of initial value also have a significant impact upon the measurement results. This paper demonstrates that a novel chameleon swarm algorithm approach could be used to ascertain temperature information from light-field multi-wavelength data at a higher accuracy level without prior emissivity knowledge. The performance of chameleon swarm algorithm was experimentally tested and compared with the traditional internal penalty function and generalized inverse matrix-exterior penalty function algorithms. Comparisons of calculation error, time, and emissivity values for each channel show that the chameleon swarm algorithm is superior in terms of both measurement accuracy and computational efficiency.

Enhanced light-field image resolution via MLA translation.

This work describes a method that effectively improves the spatial resolution of light-field images without sacrificing angular resolution. The method involves translating the microlens array (MLA) linearly in both x- and y-directions in multiple steps to achieve 4 ×, 9 ×, 16 × and 25 × spatial resolution improvements. Its effectiveness was firstly validated through simulations with synthetic light-field images, demonstrating that distinct spatial resolution increments can be achieved by shifting the MLA. An MLA-translation light-field camera was built based on an industrial light-field camera, with which detailed experimental tests were carried out on a 1951 USAF resolution chart and a calibration plate. Qualitative and quantitative results prove that MLA translations can significantly improve measurement accuracy in x- and y- directions while preserving z-direction accuracy. Finally, the MLA-translation light-field camera was used to image a MEMS chip to demonstrate that finer structures of the chip can be acquired successfully.

Spatial resolution enhancement with line-scan light-field imaging.

This Letter proposes a line-scan-based light-field imaging framework that records lines of a light-field image successively to improve its spatial resolution. In this new, to the best of our knowledge, light-field imaging method, a conventional square or hexagonal microlens array is replaced with a cylindrical one. As such, the spatial resolution along the cylindrical axis remains unaffected, but angular information is recorded in the direction perpendicular to the cylindrical axis. By sequentially capturing multiple rows of light-field images with the aid of a translation device, a high-resolution two-dimensional light-field image can then be constructed. As a proof of concept, a prototype line-scan light-field camera was built and tested with the 1951 USAF resolution chart and the high-precision calibration dot array. Good measurement accuracies in the x, y, and z directions are demonstrated and prove that line-scan light-field imaging can significantly improve spatial resolutions and could be an alternative for fast three-dimensional inspections in the production line.

Resolution analysis on light-field particle image velocimetry.

With rapid developments in light-field particle image velocimetry (LF-PIV) based on single-camera, dual-camera, and dual-camera with Scheimpflug lenses, comprehensive quantitative analysis and careful evaluation of their theoretical spatial resolutions are essential to guide their practical applications. This work presents a framework for and better understanding of the theoretical resolution distribution of various optical field cameras with different amounts and different optical settings in PIV. Based on Gaussian optics principles, a forward ray-tracing method is applied to define the spatial resolution and provides the basis of a volumetric calculation method. Such a method requires a relatively low and acceptable computational cost, and can easily be applied in dual-camera/Scheimpflug LF-PIV configuration, which has hardly been calculated and discussed previously. By varying key optical parameters such as magnification, camera separation angle, and tilt angle, a series of volume depth resolution distributions is presented and discussed. By taking advantage of volume data distributions, a universal evaluation criterion based on statistics that is suitable for all three LF-PIV configurations is hereby proposed. With such a criterion, the pros and cons of the three configurations, as well as the effects of key optical parameters, can then be quantitatively illustrated and compared, thus providing useful guidance on the configuration and optical parameter selections in practical implementations of LF-PIV.

Analysis of error propagation: from raw light-field data to depth estimation.

In micro-lens-array-based light-field imaging, the micro-lens centers serve as the origins of local micro-lens coordinate systems. Each micro-lens receives angular/depth information coded according to its center location. Therefore, the errors in positioning the micro-lens centers will lead to errors in depth estimation. This paper proposes a method that resolves error propagation from raw light-field data to depth estimation based on analyzing large amounts of simulated images with various aperture sizes, noise levels, and object distance values. The simulation employs backward ray tracing and Monte Carlo sampling to improve computational efficiency. The errors are counted and accumulated stepwise from center positioning and generation of sub-aperture images to depth estimation. The disparity errors calculated during depth estimation are shown to be more apparent either with more significant center positioning errors or with a greater defocusing distance. An experiment using an industrial light-field camera is conducted, confirming that disparity errors at considerable object distances can be reduced significantly when the micro-lens centers are positioned with higher accuracy.

Hand-eye calibration for an unfocused light-field camera.

A calibration framework is established for an unfocused light-field camera and a robotic arm. With Gaussian optics and light-field imaging principles, the mapping relationship between a point light source and its corresponding plenoptic disc feature is established, and the intrinsic and extrinsic parameters of the unfocused light-field camera are calculated through nonlinear optimization. Transformation matrices for eye-to-hand and eye-in-hand configurations are subsequently solved and are validated by applying them to an industrial light-field camera-robotic arm system. With the proposed calibration method, 3D reconstruction for calibration board in different poses is demonstrated and calibration uncertainty is discussed in detail.

Light-field multi-spectral radiation thermometry.

This Letter proposes light-field multi-spectral radiation thermometry based on an unfocused light-field camera, which can simultaneously record directions and intensities of incident rays. In this method, the direction information of rays is substituted by radiation spectrums via placing an array of filters in front of camera main lens, such that the image sensor can simultaneously acquire spectrums and intensities of rays. By decoupling a raw multi-spectral light-field (MSLF) image and utilizing traditional multi-wavelength pyrometer algorithms, the scalar field of surface temperature distribution can be achieved. To verify the method, measurement errors of different temperature levels on several typical areas of MSLF images are analyzed. In addition, the validation experiment demonstrates that accurate surface temperature measurement can be achieved with a single lens, single monochromatic image sensor, and just one snapshot in the proposed method.

Fast and Accurate 3D Measurement Based on Light-Field Camera and Deep Learning.

The precise combination of image sensor and micro-lens array enables light-field cameras to record both angular and spatial information of incoming light, therefore, one can calculate disparity and depth from one single light-field image captured by one single light-field camera. In turn, 3D models of the recorded objects can be recovered, which means a 3D measurement system can be built using a light-field camera. However, reflective and texture-less areas in light-field images have complicated conditions, making it hard to correctly calculate disparity with existing algorithms. To tackle this problem, we introduce a novel end-to-end network VommaNet to retrieve multi-scale features from reflective and texture-less regions for accurate disparity estimation. Meanwhile, our network has achieved similar or better performance in other regions for both synthetic light-field images and real-world data compared to the state-of-the-art algorithms.

Meta-Lens Particle Image Velocimetry.

Fluid flow behavior is visualized through particle image velocimetry (PIV) for understanding and studying experimental fluid dynamics. However, traditional PIV methods require multiple cameras and conventional lens systems for image acquisition to resolve multi-dimensional velocity fields. In turn, it introduces complexity to the entire system. Meta-lenses are advanced flat optical devices composed of artificial nanoantenna arrays. It can manipulate the wavefront of light with the advantages of ultrathin, compact, and no spherical aberration. Meta-lenses offer novel functionalities and promise to replace traditional optical imaging systems. Here, a binocular meta-lens PIV technique is proposed, where a pair of GaN meta-lenses are fabricated on one substrate and integrated with a imaging sensor to form a compact binocular PIV system. The meta-lens weigh only 116 mg, much lighter than commercial lenses. The 3D velocity field can be obtained by the binocular disparity and particle image displacement information of fluid flow. The measurement error of vortex-ring diameter is ≈1.25% experimentally validates via a Reynolds-number (Re) 2000 vortex-ring. This work demonstrates a new development trend for the PIV technique for rejuvenating traditional flow diagnostic tools toward a more compact, easy-to-deploy technique. It enables further miniaturization and low-power systems for portable, field-use, and space-constrained PIV applications.

Design and evaluation of a light-field multi-wavelength pyrometer.

This letter describes the design and implementation of a multi-wavelength light-field pyrometer, where six-channel radiation images were captured with one CMOS sensor. Such capability is achieved by placing a 2 × 3 filter array in front of the main lens of an unfocused light-field camera, such that discrete wavelength and radiation intensity can be simultaneously recorded. It demonstrates that through black-body furnace experiments, how multi-channel radiation images can be extracted from one raw light-field multispectral image, and how accurate 2D temperature distribution can be recovered by optimization algorithms.