This PDF file contains the front matter associated with SPIE Proceedings Volume 12607, including the Title Page, Copyright information, Table of Contents, and Conference Committee information.

The increasing demands in industry, for example for products in the consumer electronics sector or for assistance systems in cars, and the continuous development in semiconductor are leading to significant miniaturization in electronic components. These requirements are inevitably also transferred to ultra-precise manufacturing and thus ask for monitored production steps. In the context of Industry 4.0 and other developments in the context of modern production sensors to supervise production steps are crucial. An essential component here is non-destructive testing (NDT) and specifically optical metrology. Complete 3D measurement of objects using white light interference in mass production is time consuming and therefore not well-suited for use as a means of inline inspection. In addition, the long measurement cycles with complex sequences generate large amounts of data and the effort for processing the data must also be included in overall considerations. The Flying Spot Scanner (FSS) intelligently avoids these disadvantages and is therefore ideal for inline inspections. To meet the need of additional, even complex and time sensitive measurement tasks, Precitec Optronik has developed the Flying Spot Sensor. The active measuring head was specially developed for in-line use and ideally complements the spectral interferometric sensor to form a smart inspection system. The light from the sensor is coupled into the measuring head via a light guide and deflected by a mirror system, a so-called galvanometer scanner. Finally, the light passes through a telecentric lens, which serves as a focusing module on the outward path and as a measuring aperture for the reflected light. Due to the movable mirror system, the measuring light beam can be deflected at different angles and thus the measuring spot can be freely positioned within the field of view of the lens. Long paths of linear axes are replaced by short rotary movements, resulting in an extreme reduction of the measuring, or scanning time. By using specialize focusing modules, the Flying Spot Scanner can be adapted to different application scenarios. These optics are characterized by a low curvature of the focal plane, very small telecentric errors and a very large depth of field. The measurement system can also be used in two operating modes, a thickness mode, or a distance mode. The two operating modes can be selected at will via the digital interface, which means that the switching process can be easily integrated into an automatic measuring sequence.

This article presents off-axis modulation transfer function (MTF) measurements in the image field covered by the azimuthal angle θ = ±20 deg and a defocus range of Δz ∈ [−40 μm, +40 μm] with the slit based MTF measurement setup at PTB. The experimentally determined standard deviation of a set of N = 10 MTF repetition measurements under a certain set of repeatability conditions is employed to estimate the standard uncertainty uexp with respect to the repeatability of measurements in the setup. The importance of measurement repeatability and measurement reproducibility in the assessment of measurement accuracy and measurement uncertainty for interlaboratory comparisons is discussed.

Methods of using optical frequency comb for angle measurement are discussed. Autocollimation is the most commonly used method for angular displacement measurement. Here, a diffraction grating is attached to the measurement target instead of a mirror, which is used in the conventional autocollimation method. Each mode of the optical frequency comb with a wide wavelength bandwidth is angularly dispersed by the grating. The angular displacement of the target is determined from the most strongly detected optical frequency modes.

We report real-time analysis of flue gas parameters in a power plant boiler by a broadband supercontinuum lidar system. The system utilizes differential absorption between three wavelength bands at 1250 – 1500 nm to simultaneous measure fly ash aerosol particle distribution and water vapor temperature and concentration profiles using a single measurement port.

We investigate using machine learning techniques to infer various physical properties of rocks from hyperspectral imaging data. In particular, we demonstrate that deep neural networks (DNN) can infer mechanical and geochemical properties based on high-resolution Fourier transform spectrograms. Our goal is to enable real-time petrophysical analysis of subsurface rocks and fluids. The ongoing work encompasses the development of sensors and algorithms that facilitate non-destructive, fast, and high-resolution mapping of petrophysical properties. We acquired high-resolution mappings of mechanical, chemical, and electromagnetic properties at sub-millimeter scale (> 100 um) using a scanning system fitted with multiphysics probes, including impulse hammer geomechanical probe designed to measure the rebound hardness and the reduced Young modulus; Fourier transform spectrometer (FTIR) to acquire the diffuse reflectance; acoustic transducers to measure unconstrained sonic velocities; and near-surface gas permeability. We have characterized over two hundred thousand samples across various lithologies, including limestone, sandstones, and shales from outcrops and cores from unidentified wells. We present the results of machine-learning models and algorithms that predict, based on the IR reflectance data, the rock types, and the unconstrained geomechanical properties. The method could be extended to characterize other solids from subsurface, terrestrial, or non-terrestrial environments. The combination of photonic measurements and machine learning provides the means to find non-causal relations between materials' electromagnetic/photonic response and their other physical properties under various stress states and environmental configurations. This work presents the foundational blocks to achieve this objective and develop optical sensors for sustainable energy extraction.

Due to legal changes, the use of agents for soil disinfection in arboriculture is severely limited. For this reason, tree nurseries must look for new ways in which weed control can be carried out on seed beds in the future. In cooperation with the local Chamber of Agriculture we want to explore the use of Unmanned Aerial Vehicles (UAVs) for the autonomous weed detection and weed management in arboriculture to develop an innovative approach for weed control. UAVs are selected because of the disadvantage of ground-based robotic systems in agriculture which is the limitation of their usability in wet soils, as most robotic systems get stuck and are no longer manoeuvrable. Climate change is causing these weather conditions to occur more frequently. This paper evaluates early research results of a vision-based crop row detection system for UAV-based weed detection in arboriculture in regards to overall detection performance, real-time capability, weight, size and power-consumption.

The significant development of terahertz wave technology requires precise measurement of terahertz optical devices such as diffractive gratings with micrometer-scale periodicity. We propose a new measurement method for fast, robust and precise shape measurement of micro-periodic structures, which can be regarded as a scan-less version of the deflectometry. Whereas the deflectometry demands the scanning of the beam spot in order to collect the tilt angle information from various different position on the sample surface, the proposed method simultaneously obtains it from a single diffraction image, then reconstructs the sample shape based on a light reflection model called ray reflection model. In comparison to the interferometry, the proposed method is principally robust to the mechanical vibration because the diffraction image is hardly affected by the displacement of the sample. The limitation of the proposed method is also discussed, and the mathematical expression of the constraint conditions required for the shape reconstruction is clarified. The numerical experiment based on the electromagnetic simulation with rigorous coupled-wave analysis (RCWA) demonstrates the possible accuracy on the order of 10 nm and the effectiveness of the use of the incoherent light. The physical experiment is also conducted by the constructed optical system, and the fundamental validity of the measurement result of the proposed method is confirmed.

Diameter is one of the most fundamental and important parameters that characterize the optical properties of a tapered fiber, so it is necessary to accurately measure its diameter. In this study, we proposed a method for measuring the diameter in the sub-1 µm diameter region of a tapered fiber by measuring the spatial period of the standing wave formed by counter-propagating light waves incident from its both sides. We used a scanning near-field optical microscopy (SNOM) probe fabricated from an optical fiber to measure the standing wave intensity distribution along the tapered fiber axial direction and its spatial period, from which the tapered fiber diameter can be estimated.

Due to legal changes in soil disinfection and agrochemical application, using soil disinfection agents and other agents is increasingly restricted. This restriction and more frequent extreme weather events pose new challenges for horticultural production. Tree nurseries are looking for new ways of weed management for their seedbeds. Autonomous smart agricultural robots have the potential to support or replace chemical weed management as well as labor and cost-intensive manual weeding, contributing to the sustainable transformation and optimization of agricultural production systems. A significant challenge for conventional ground-based autonomous systems is working in environments more often inaccessible to heavy machinery because of rain-soaked soils. Unmanned aerial vehicles (UAVs) are one of the most promising technologies for weed detection and management because they are not constrained by ground conditions, can be easily maneuvered, and can cover a large agricultural area. This paper presents the design of a low-cost parallel delta robot for mechanical weed control in horticulture that uses a machine-vision system for weed detection. The system is designed to be carried by both conventional and UAV-based autonomous vehicles. Moreover, an evaluation of the positioning accuracy is presented using an early prototype for a conventional carrier system.

In the conventional deflectometry, since the observation target is a virtual image generated by the mirror, measuring is highly sensitive to the titling angle of mirror. That means it is hard to use this method to measure the displacement. Here, we proposed a new type of deflectometry method, in which the observation target is the surface of the observed mirror. Using an imaging lens, the measurement fringes are projected on the surface of mirror. The result show that the proposed method is lower sensitive to the titling angle of the mirror compared to the conventional deflectometry.

Unmanned aerial vehicles (UAVs) can overcome several limitations of satellite and aerial platforms using their multiple visit ability. However, UAVs usually collect images of small and simple regions from a large image scene and obtain high-resolution images from various viewing angles and altitudes. Multiple datasets created in various regions and conditions can be helpful considering data expansion to improve the usability of the UAV datasets with deep learning. The combined segmentation network (CSN), which can train two datasets simultaneously by sharing encoding blocks, was used to segment heterogeneous UAV datasets, such as UAVid and semantic drone dataset. CSN shared encoding blocks to learn general features from two datasets and decoding blocks trained separately on each dataset. For the preprocessing step, classes of each dataset were adjusted to minimize the difference between the two datasets. Experiment results show that CSN can segment more accurately for specific classes, such as background and vegetation, which have low ratios in the single dataset. This study presented the potential application of integrated heterogeneous UAV imagery datasets by learning shared layers. Thus, surface inspection would be effectively conducted using UAV datasets.

Laser-based ultrasound and laser vibrometry are quite mature technologies that have been in use in both manufacturing applications and applied research for several decades. One area of continuing research is in the field of compensated vibrometry, where shot-noise limited performance is sought under real-world manufacturing conditions, including in situ inspection diagnostics and process control of such in-factory applications as additive manufacturing and thermal processing. We propose a compensated vibrometry system, involving a novel spatial light modulator, comprised of a metasurface retroreflector array, combined with a MEMS spatial phase modulator, with the potential to achieve adaptive optical correction of dynamic workpiece distortions, leading to shot-noise limited detection of ultrasound surface displacements and vibrations.

An optical machine system for measuring the depth distance of an object was proposed. This system can scan a region by obtaining the relationship curve between focus distance and motor step. The depth distance was acquired by determining whether the captured objects was in focus or not by discerning the sharpness of the object’s outline, with the motor step when the object was in focus. This system has been verified its feasibility in the experiment and has the merits of a larger sensing range, lower manufacturing cost, and can be used outdoors.

According to estimates by the German Wildlife Foundation, approximately 90,000 fawns are injured or killed during harvesting operations in Germany each year. In recent years, drones with thermal imaging cameras have been increasingly used to locate fawns before harvesting and bring them to safety. However, this process is often time consuming and requires several people to conduct a search. In this paper an autonomous fawn tracking system based on drone images and CNNs is presented that automatically performs a fawn search with a high positioning accuracy.

While many instruments for the measurement of optical surfaces are based on electronic cameras with a regular sampling grid given by the pixels of the image sensor, there is a class of measurement instruments that use pointwise data sampling, either with tactile or with optical sensors. In this case a variety of sampling strategies exists. In our presentation we show that the sampling strategy has a big influence on the resulting measurement uncertainty. For the measurement of aspheric optical surfaces we describe a procedure to select an optimized strategy, extracting a maximum of information from a minimum of sampling points.

The research proposal was used hafnium dioxide (HfO₂) and silicon dioxide (SiO₂) as the high and low refractive index for the multilayer anti-reflection (AR) films were deposited on a flexible polyethylene terephthalate (PET) by electron-beam evaporator with ion-beam assisted deposition (IAD). The optical and stress properties of these multilayer (HfO₂/SiO₂)² films were investigated. A homemade phase-shifting shadow moiré interferometer was used the automatic measurement system to catch the interferograms and calculated the residual stress of flexible electronics. The experimental results show the optimal oxygen flow of HfO₂ and SiO₂ were 15 and 25 sccm, respectively, for the multilayer anti-reflection coating (AR coating) with electron-beam evaporator. The 1st and 2nd layers have light thickness resulted from low ion-assisted energy, so the stress generation is very low. Since the compressive stress is not generated until the 3rd and 4th layers. However, the stress is reduced after high and low refractive index stacking, and the final residual stress of (HfO₂/SiO₂)² multilayers on the PET is -473.37 MPa.

We will present a new approach of the linearized focal plane technique (LIFT), formerly developed by ONERA, which results in an improvement of a factor of 16 (4x4) of the spatial resolution. This technology is based on the combination of standard SH technology with phase retrieval algorithms applied on all spots of the microlens array that provides information on high spatial frequencies. We will show some measurements performed on extremely complex wavefronts. This technology presents very promising perspectives for optical and freeform metrology and can advantageously replace, at lower cost and better usability, Fizeau interferometry.

We will present an innovative method for the measurement of Parallel Optics or optics with parallel surfaces, often planes such as windows, filters, and mirrors, but also potentially spherical in the case of domes. This new and patented approach solves issues related to the metrology of this type of sample, mainly associated with the signal reflected by the back surface of the substrate. Instead, and with no extra hardware or specific optical add-on making the testing more complicated and expensive, our implementation allows for a straightforward characterization of Parallel Optics with no manipulation of the sample, nor preparation of any kind. We will present the concept as well as its implementation and results obtained on samples as compared with other historic reference techniques.

An octuple-pass configuration for reflectance confocal system is demonstrated. By utilizing the lens in quadrantal division as well as retro-reflection approach, the same propagating beam is configured to pass the lens eight times and directed to hit the same scanning point on the test surface as much as four times. With the increased number of passes, it is shown that the axial resolution of octuple-pass system is four times better than that of the conventional double-pass system. The effects of the beam size and its incident orientation on the lens are also investigated.

Optical constants and thickness of a single layer on the transparent substrate can be extracted simultaneously from four ellipsometric parameters -Ψ_𝑟 , Δ_𝑟 , Ψ_𝑡 , Δ_𝑡 -, which are obtained through reflection and transmission ellipsometry measurement at a single wavelength and a single incidence angle. A transparent substrate, however, induces the problem of backside reflections, and then incoherent superposition of light. In this work, the effects of such reflections were empirically investigated as a function of incident angle and analyzed by using single-point measurement results obtained with a commercialized spectroscopic ellipsometer. Our study shows that at the Brewster angle illumination, the effect of backside reflection can be minimized. Based on this result, a reflection and transmission imaging ellipsometer was configured for the measurement of a Si film deposited on a quartz glass plate. The measurement results showed the availability, as well as the limitation, of the reflection and transmission imaging ellipsometry.

A double-reflection confocal probe technique for surface profiling microscopy is presented. By delivering a small collimated laser beam at an off-axis position into a microscope objective to allow retro-reflection on the first reflected beam from the surface under test, second reflection at the same point of the test surface can be established. The feasibility of this technique is experimental investigated. This approach improves the axial resolution by a factor of two as compared to that of the conventional single-reflection system. Simulation is performed to estimate and evaluate the lateral scanning performance of the double-reflection system.

This paper proposes a process for generation of structural coloring without paint, involving the production of nano-scale porous shapes on the surface of aluminum alloys and sputtering with platinum particles. The first step is anodization, with micropores of 10 to 40 nm in diameter and oxide films of 71 to 1,151 nm in thickness formed on the surface. Coloration density was found to vary with sputtering times between 1 and 4 minutes. The results of the study showed that the process can be applied to generate a wide range of colors within the visible spectrum.

Recently, there has been a high demand for high-quality imaging from weak light in measuring micro-defects. Deep Learning Ghost Imaging (DLGI) has been proposed as a fast and sensitive imaging method for defect inspection. However, measurement with deep learning has a problem evaluating the prediction uncertainty. The predicted value from deep learning is distributed close to the true value in the feature, while the traditional measurement value is physically distributed close to the true value. Then, applying the conventional uncertainty evaluation method based on statistics is difficult. To overcome this problem, we propose the evaluation method of the prediction uncertainty based on the feature map in the middle layer of the CNN. By adding random numbers to the middle layer, several close estimates of feature values can be obtained. The standard deviation of these estimates is defined as prediction uncertainty. This paper shows the numerical comparison of the proposed method with evaluation by data augmentation, which evaluates the prediction uncertainty by adding fluctuations to the input data. The data augmentation method can estimate the uncertainty of changes in measurement conditions. Although the data augmentation method does not provide enough change for low SNR data, which makes uncertainty evaluation difficult, the proposed method offers constant fluctuation even for low SNR data. We have numerically confirmed that the proposed method can accurately evaluate the prediction uncertainty even for low SNR.

A luminous target marker was developed to measure motion errors of machine tools by motion capturing. The target marker consists of a polyacetal diffuser ball and high intensity LED light. The influence of the marker intensity on the fluctuation in the measured position was investigated. The negative correlation between the fluctuation amplitude and marker intensity was observed. Thus, the marker intensity should be enough high to obtain the uniform intensity distribution and decrease the fluctuation in the identified position. The measurement of circular motion trajectory was also demonstrated and compared to the measurement by a grid encoder that is conventionally used in two dimensional motion error evaluation of machine tools. The trajectory measured by the motion capturing was similar to that by the grid encoder in micrometer level. The quadrant glitch that is a typical dynamic motion error in machine tools was measured by the motion capturing even with 100 mm defocus. The experimental results showed the capability of motion error measurement with quick and easy installation by the motion capturing system.