For the evaluation of the surface topography only a small area of the surface in comparison to the whole workpiece is measured. Therefore, markings are needed as reference to relocate the measurement area. For the comparison of different optical or tactile measurement devices, Frühauf et al. manufactured silicon reference samples with etched marks [1]. Another application are before and after measurements to visualize and evaluate the change related to manufacturing processes. Newton et al. used micro-milled slots as relocation landmarks [2]. There is no description if and how they protected those landmarks during manufacturing. Moretti et al. describe algorithms for relocation based on landmarks that can be found both on the original and the manufactured surface [3]. This method can assess partial changes, since the landmarks of the original surface are needed. The relocation under the microscope is done by visual estimation. A proper reference is missing. With this paper new approaches with permanent abrasive markings for those applications are presented. A microscope (Confovis Duo Vario) executes the surface topography measurement. It combines confocal mode, that is used here, and focus variation mode. The surface data is processed with MountainsMap® 7. For the comparison of different optical measurement devices markings with the needle of a height marker are used on a milled surface (Figure 1a and b). In addition, a laser labels the different measurement areas (Figure 1a). The needle lines are used for positioning. They are excluded from the measurement. MountainsMap® provides the function colocalization to superimpose the different data (Figure 1c). Then profiles are extracted and compared to show similar behaviour (Figure 1d). Both devices generate comparable results. The second application are before and after measurements for a blasting process. As markings, Vickers hardness indentations are impressed. A 3D-printed cover protects the area with markings during the manufacturing process (Figure 2a, b and c). After blasting the before and after measurements are superimposed with the use of the indentations. Then profiles are extracted. The profiles show the change created by the impact of the blast media (Figure 2e). The paper will present the comparison by calculating differences for example in volume. As another example, the blasting of additive manufactured surfaces is analysed. The solution for markings depends on the application. The examined markings provide landmarks for proper relocation during the measurement and data evaluation.

Transportation-related air pollution, particularly fine particulate matter (PM) from non-exhaust sources, presents significant health and environmental challenges. Roadside vegetation has been recognized for its role in mitigating urban air pollution through the utilization of leaf attributes like surface roughness and trichomes, acting as natural PM filters [1, 2, 3, 4]. Despite potential variations in particle capture among plant species, a comprehensive understanding of the underlying mechanisms is lacking [5]. This study employs a two-phase approach, first by the identification of leaf characteristics that promote the retention of PM generated by tire wear. Subsequently, diverse plant species are chosen based on their attributes, intended to serve as barriers against the dispersion of PM. To simulate exposure to the PM source, a laboratory analytical setup was developed. Four well-adapted leaves from roadside vegetation in France were chosen, representing diverse surfaces with distinct retention mechanisms: Vanhouttei de spirala (1), Euonymus Europaeus (2), Cotoneaster aspressus (3), Cupressus leylandii (4). Using a binocular microscope and 3D optical profilometry, PM capture mechanisms and leaf surface characteristics were analyzed, identifying effective PM removal options among different plant species. This approach provided insights into particle-leaf surface interactions, evaluating size-dependent capture efficiency for effective PM removal. Observations revealed that both leaf topography and particle characteristics significantly influenced particle retention (Figure 1). Specifically, leaf 2 did not retain tire particles in the same manner as soil and aggregate particles, which were effectively captured by leaf 4. The study establishes correlations between leaf micromorphological features and pollutant capture, contributing to the design of effective plant barriers for air decontamination. In conclusion, it underscores the global importance of addressing health impacts resulting from non-exhaust PM emissions, emphasizing the necessity for proactive planning and interdisciplinary collaboration to implement efficient green infrastructure solutions.

Characterizing surface anisotropy is an important aspect of surface metrology considering it can be studied after a manufacturing process as quality control but also to predict and optimize the surface functionality. It helps researchers to understand how surface topography vary in different directions, which is essential for optimizing material performance and ensuring reliability in diverse applications, from industrial to medical. Surface anisotropy, being a parameter that changes with scale, requires a multiscale [1] approach to avoid biasing the results. However, conventional methods using the standard parameters as in ISO 25178-2 [2] are only relevant at a specific scale. The previous advancement in multiscale curvature characterization involved using curvature tensors to determine the most prominent directions for ridges and valleys [3]. That method allowed both quantifying and visualizing surface anisotropy. However, the concept of anisotropy can appear holistic when considering that a surface may exhibit various types of features. For instance, a milled and sandblasted surface or the skin texture, encompass different types of features. Our approach now is to isolate these features to create visualisation of directionalities and geometric characterizations in order to gain a better understanding surface anisotropy. Considering the curvature distribution for a given scale of observation distinctive modes can be considered as candidates for further extraction. For example, a flat surface with bidirectional groove pattern should produce evident peaks in the maximum curvature distribution for a scale related the size of these grooves. That peak should correspond to the inverse value of radii of the grooves. Adding the information about the maximum curvature direction, unidirectional grooves can be further extracted. Using color-coding features of similar morphologies can be visualized on the rendering of surface topography. In this study, we use canvass surfaces painted by both left-handed and right-handed individuals (Fig. 1a). These surfaces exhibit various trends and a strong anisotropy. The primary purpose of this experimentation is to seek relevant parameters to differentiate surface texture with the factor of handedness. This type of morphology exhibits features with a directional tendency. The valleys extracted at the scale of 40× original sampling interval are depicted in Figure 1b. The proposed method should contribute to better understanding the nature, sizes, and direction of surface features.

In optical metrology, the growing applications of surface texture characterisation contribute to an increasing need for advanced measurement technologies [1]. Coherence scanning interferometry (CSI) is a popular technique in three-dimensional microscopy offering nanometre-level height resolution [2]. Various optical specifications and illumination features influence the accuracy of the surface topography measurement systems. In CSI, the bandwidth of the light source and the numerical aperture of the objective lens change the width of the coherence envelope and/or fringes along the z-axis [3], consequently affecting the accuracy of the measured surface topography. Moreover, the polarisation of the incident light plays an important role in topography measurement using CSI. The polarisation mode of the light can affect how it interacts with surface features. Certain surface structures may exhibit varying reflective or scattering properties based on the polarisation of the incident light. The effect of polarisation on the thickness and optical properties of thin films [4] and the batwing effect due to diffraction from sharp edges on the surface has been studied [5, 6]. The Boundary element method (BEM) is an electromagnetic light field scattering model that provides a rigorous solution to Maxwell’s equations [7]. The BEM model can predict the scattered light from diverse surface geometries, including complex features such as overhangs and undercuts, while accounting for the effect of the polarisation of the incident light [8]. In this study, BEM is used to model CSI (BEM-CSI) with a mean wavelength of 0.6 µm, full-width at half maximum wavelength bandwidth of 0.08 µm and a numerical aperture (NA) of 0.2. The effect of the polarisation of the incident light is investigated on the simulated interference signal corresponding to a silicon sinusoidal profile with a period of 4 µm, peak-to-valley of 0.2 µm and a complex refractive index of 4.289 + 0.0485i (see the red curve in Figure 1) using the BEM-CSI model. Figure 1 shows the cross-sectional view of the interference fringes in the xz-plane obtained by the BEM-CSI model for the (a) transverse electric (TE) and (b) transverse magnetic (TM) modes of the incident light. As depicted in Figure 1, various polarisation modes can introduce a phase shift in the interference pattern leading to errors in reconstructing the surface topography when employing a phase-dependant algorithm, such as Fourier domain analysis (FDA) [9]. Moreover, the visibility of the interference signal depends on the polarisation, thereby influencing the accuracy of the results. Figure 1 (a') and (b') illustrate the reconstructed profiles obtained by the interference pattern of Figure 1 (a) and (b) using the FDA algorithm. Figure 1 clarifies the necessity of choosing an appropriate polarisation to ensure accurate and reliable surface measurement results using a CSI instrument, particularly when dealing with materials or surface features that exhibit polarisation-dependent behaviours. In this study, the effect of polarisation on various surface topographies, and surface materials considering different NAs will be discussed.

In precision engineering, particularly in the context of micro gear production, stringent tolerance levels are essential for ensuring optimal functionality. Traditional tolerance design models in this field tend to focus primarily on product value, often overlooking aspects such as product robustness, which presents challenges in maintaining a balance between manufacturing precision and cost efficiency. This study proposes a hybrid optimization method that combines the predictive strengths of Bayesian techniques with the comprehensive search capabilities of genetic algorithms. This method is designed to address both precision and cost-effectiveness in micro gear manufacturing. The research methodology employed integrates statistical analysis with algorithmic modeling. Bayesian methods are utilized for forecasting and adjusting tolerance levels, using historical data and probabilistic models to enhance manufacturing accuracy. Concurrently, genetic algorithms are applied to explore a range of design parameters, aiming to identify optimal tolerance settings. This approach is known for its effectiveness in tackling complex optimization problems. The combination of these methods allows for an extensive exploration of potential solutions, seeking to achieve an equilibrium between precision and cost. Early results from this study suggest an improvement in tolerance optimization compared to traditional single-method approaches. The validation of this approach was conducted on a dataset comprising real-world micro gear manufacturing scenarios. The findings indicate that this method is applicable and effective in real-world manufacturing settings, offering a feasible approach to balancing precision and cost in micro gear production.

In the manufacturing industry, metrology assumes a critical role in affirming the quality of end products. [1] In cases where specific precision components fail to meet quality specifications, they are often scrapped, posing sustainability challenges for the manufacturing process. With the help of AI and smart sensors, we can move towards zero-defect manufacturing with an explainable approach. [2,3] Turn-mill machining center DMG is used to perform the cutting experiments. Apart from the machining parameters, a 3-axis dynamometer is used to record the cutting force. A chromatic confocal sensor was used to record the roughness as ground truth for on-machine measurement of roughness. 200 different cutting conditions were used to design experiments consisting of up and down milling. Instead of implicit feature extraction procedure with deep learning-based models, the work proposes a force lobe- based approach by creating windows within the force signals. Further different statistical features such as mean, standard deviation, skewness, kurtosis, crest factor etc were extracted from these windowed signals. Furtsssher these features were given as input to Machine learning model. Multiple ML models were tested but Random Forest was chosen as the base model because of its implicit explainability and overfitting handling criteria. Further, by experimenting with combinations across predefined parameter grids, grid search is used to methodically investigate and adjust hyperparameters for machine learning models, guaranteeing optimal performance. A regression model was developed to predict the roughness and a classification model was used to predict whether the signals belong to the category of Up or down milling based on the cutting force signals. The results show a strong correlation between the lobe-based features with the final output of prediction of surface roughness. This is validated from the waterfall graph showing the feature importance for the prediction models. This work is also the starting point of a proactive quality control model focusing on the process monitoring part.

In the domain of smart manufacturing, precise temperature monitoring is crucial, particularly for assets like arch, kiln, or arc furnaces in industries such as foundry and glass production. Traditional methods like thermocouples offer only sporadic and isolated measurements. To address this limitation, recent advancements in Optical Time Domain Reflectometry (OTDR) have enabled distributed temperature measurements. However, the applicability of optical fibers is constrained to temperatures below 700°C, due to durability concerns over extended industrial usage. This paper presents a comprehensive study on the deployment of Optical Fiber and Electrical Time Domain Reflectometry (ETDR) technologies, tailored for high-temperature monitoring in industrial environments. The focus is on the development of robust sensors capable of enduring extreme conditions, particularly temperatures above 700°C, where OTDR sensors reach their limits. For temperatures below 700°C, the utilization of optical fibers presents a viable solution. The study elaborates on the advancements in optical fiber technology, particularly in enhancing their temperature endurance and measurement accuracy, making them suitable for a wide range of industrial applications. In the context of higher temperatures, ETDR sensors emerge as a promising alternative. The study introduces an innovative ETDR-based sensor system, capable of enduring and accurately measuring temperatures beyond the threshold of optical fibers. This system employs a high-frequency electro-magnetic signal within a transmission line composed of materials specifically chosen for high temperature and durability. The backscattered signal analysis from these ETDR sensors provides a distributed and continuous temperature profile along the sensor line. A significant portion of the study is dedicated to material selection and aging properties to optimize the response of ETDR sensors to high temperatures. Various materials, including alumina as a dielectric and different metals as conductors, are investigated for their suitability in ETDR systems. The paper discusses the experimental and modeling approaches employed to identify the most efficient material combinations and sensor geometries, considering factors like permittivity, resistivity, permeability, and environmental influences such as temperature, humidity, and mechanical strain. The paper aims to promote the development of these new technologies, offering a pathway towards more efficient and reliable temperature monitoring in smart manufacturing settings. By integrating optical fiber and ETDR technologies, the study paves the way for enhanced monitoring capabilities, contributing to the safety, efficiency, and longevity of various industrial processes. Main References [1] F.-K. Chang, Structural health monitoring 2000. CRC Press, 1999. [2] J. A. Stastny, C. A. Rogers, and C. Liang, "Distributed electrical time domain reflectometry (ETDR) structural sensors: design models and proof-of-concept experiments," in Smart Structures and Materials 1993: Smart Sensing, Processing, and Instrumentation, 1993, vol. 1918: International Society for Optics and Photonics, pp. 366-376. [3] R. H. Cole, "Time domain reflectometry," Annual review of physical chemistry, vol. 28, no. 1, pp. 283-300, 1977. [4] S. B. Jones, J. M. Wraith, and D. Or, "Time domain reflectometry measurement principles and applications," Hydrological processes, vol. 16, no. 1, pp. 141-153, 2002. [5] B. M. Lee, K. J. Loh, and F. L. di Scalea, "Distributed Strain Sensing Using Electrical Time Domain Reflectometry With Nanocomposites," IEEE Sensors Journal, vol. 18, no. 23, pp. 9515-9525, 2018. [6] S. Sun et al., "A novel TDR-based coaxial cable sensor for crack/strain sensing in reinforced concrete structures," IEEE Transactions on Instrumentation and Measurement, vol. 58, no. 8, pp. 2714 2725, 2009. [7] T. Wei, S. Wu, J. Huang, H. Xiao, and J. Fan, "Coaxial cable Bragg grating," Applied Physics Letters, vol. 99, no. 11, p. 113517, 2011. [8] Z. Zhou, T. Jiao, P. Zhao, J. Liu, and H. Xiao, "Development of a distributed crack sensor using coaxial cable," Sensors, vol. 16, no. 8, p. 1198, 2016.

Femtosecond lasers offer significant advantage in precision manufacturing due to the ability to create surface textures with a minimal heat affected zone avoiding undesirable bulges that can alter the functionality of the surface in applications such as tribology [1]. The selection of optimal parameters for achieving bulge free textures remains a complex task based on trial-and-error approaches. Current heat accumulation prediction models assume that the unused ablation energy remains in the material as residual heat, leading to thermal damage prediction [2]. Nevertheless, these models do not predict bulge generation. This study presents a new semiempirical model for bulge generation prediction in multi-shot dimple and line ablation. The model is based on the mean energy delivered per unit area throughout the ablation process, i.e. the accumulated fluence, and validated through the measurement of the generated bulges. A femtosecond laser was used to ablate multi-shot dimples and lines with various process parameters combinations. Sensofar 3D optical profiler employing confocal technology with EPI100X objective was used to measure the generated geometry, and bulge’s mean and maximum height and volume were characterized (see Figure 1). After analysing the evolution of the mean and maximum height and volume of the bulges with the accumulated fluence, a relationship with the bulge’s volume was found as observed in Figure 2 (a). Results indicated two regimes: (i) absence of bulge generation and (ii) bulge generation, with logarithmic increase of the normalized volume as the accumulated fluence rise (see Figure 2 (b)). This trend was consistent for both dimples and lines with limits for bulge formation calculated at 31.48 ± 8.8 J/cm2 and 30.67 ± 13.7 J/cm2 respectively. The study concludes that accumulated fluence serves as a viable metric for establishing a limit for bulge generation. Incorporating the limit of accumulated fluence into existing prediction models for laser ablation profiles represents an improvement as it guarantees the selection of process parameters that create textures with bulge-free edges without the need of trial and error.

To optimally deploy optical instrumentation more widely within manufacturing chains and aid a move towards Industry 4.0, sensors need to be developed in a form more suitable for integration [1]. Current optical instrumentation is often too bulky and heavy to perform in-process and on-machine measurements within manufacturing systems without interfering with ongoing processes. One reason why current instrumentation is too large is the number and size of the optical components, such as glass lenses, utilised in their construction. A reduction in the size or number of these elements would allow significantly more compact optical instruments to be created. One instrument we have particular interest in is a Single-shot Dispersive Profile Interferometer (SDPI) [2], which is designed to take on-machine measurements of the surfaces of moving substrates in roll-to-roll processes for printed electronics. In an SDPI, light from the image of a line on the measurand surface and that from a reference beam is combined before being separated out by wavelength in a direction perpendicular to the measurement line, producing a spectrum for each point along it. The variation of intensity within the spectrum from each point encodes the height of the surface at that location, allowing a profile to be determined from a single image. The determination of the spectrum at each point along the line is currently carried out using a traditional form of spectrometer, one that requires multiple components to separate the different wavelengths of light and focus them onto different regions on an areal CCD/CMOS detector. However, in this form the spectrometer takes up a significant portion of the size of the overall instrument which limits it potential for integration. Photonic metasurfaces are an emerging technology which allows the manipulation of light without the bulk of traditional optical elements, with elements having a thickness on the order of the wavelength of light. Furthermore, they can be designed to perform multiple functions simultaneously, reducing the need for the multiple optical elements required in many applications [3]. Here we present our work towards developing a single metasurface element that combines focussing and wavelength separation functions, allowing a spectrometer to be formed with just the addition of a detector [4]. This approach can significantly reduce the size of the spectrometer, lowering the number of elements that need to be aligned (and maintained in alignment). The metasurfaces utilised here uses the truncated waveguide approach and consists of an array of GaN nanopillars fabricated on an Al_2 O_3 substrate. By varying the form of the pillars across the surface, the phase delay imparted on the light can be changed by up to 2π radians, and in this way incident wavefronts can be reshaped arbitrarily as they pass through. The spectrometer harnesses the strong chromatic aberration that is present with basic metalenses of this form in order to angularly separate the light. The metasurface is designed to take light from a specific point, and at a specific wavelength, on the object side of the metasurface and focus it to a point on the detector side. The chromatic aberration present leads to a shift of the focal point on the detector side for other wavelengths (see figure 1). A shift of the point source in the y-direction corresponds to a shift where the spectra is formed. In this way the spectrometer required by the SDPI is formed. Here we will present details on the design, fabrication, and characterisation of the metasurface-based spectrometer, which is a crucial stepping stone to developing a more compact SDPI instrument in the near future.

In this work, we propose a data fusion pipeline for registering two point clouds that have a distinct disparity in the size and point density: one point cloud is large and sparse, while the other is small and dense. The small and dense point cloud represents the surface texture details of a region located within the large and sparse point cloud. The motivation behind the data fusion pipeline proposed in this work is that most of the latest data fusion algorithms are only applicable to point clouds with similar sizes and point densities. In this algorithmic pipeline, firstly, the large and sparse point cloud is defined as the reference and it is segmented into subsections (“sub-clouds”), each of which is sized identically to the small and dense point cloud. Then, the distances between points and the plane formed by two of the main axes (principal component analysis – PCA) are calculated and plotted into histograms for the small and dense point clouds and each sub-cloud. Sub-clouds showing similar histograms to the small point cloud’s histogram are selected as candidates. In the next step, the angles between local normal vectors and the PCA-plane are calculated and plotted into histograms for the small and dense point cloud and all candidate sub-clouds. The sub-cloud generating the most similar histogram to the small point cloud’s histogram is determined as the best matching. The small point cloud is then transferred into the coordinate frame of the large point cloud and translated to the target sub-cloud’s location. The next step is to determine the correct orientation of the small point cloud: the small point cloud is rotated around 360° in 10° intervals and voxelised into (10 × 10) voxels on the x-y plane and five layers along the z-axis (i.e. 500 voxels in total for each point cloud). The target sub-cloud is voxelised simultaneously with the same configuration. The percentage of points falling in each voxel (i.e. the percentage relative to the total number of points in the processed point cloud) is then calculated. The voxels of the small point clouds and each candidate sub-clouds are compared : if two corresponding voxels for these two point clouds have similar percentages, this pair is defined as a good match. Finally, the orientation of the small point cloud which gives the largest number of matched voxel pairs is determined as the correct orientation for registration. The proposed pipeline has been tested on synthetic three-dimensional models characterised by artificial engineered surfaces.

The Textures are becoming more and more key in the perceived value of our packaging, in the adequation between formula and packaging and for our consumers experience. Our objectives in terms of sustainability and recyclability lead us to reduce the diversity of type of materials that we can use but also we need to increase the diversity of sensorial tactile textures and consumer experience.The diversity of textures achievable on a packaging is also a large territory of technical improvements. All these needs create the necessity of having an objective and quantitative way to characterize and evaluate a texture and its corresponding tactile perception. This work aims to develop new knowledge and establish a multisensory database (roughness, touch) to facilitate the substitution of conventional plastic packaging solutions with more environmentally friendly alternatives, while reducing post-processing steps (varnishes, paints) that negatively impact the recyclability of products. We conducted a cross-study on the roughness and touch of materials used in packaging manufacturing. The goal was to establish a classification of the tactile sensorial properties of the samples. Initially, we analyzed the surface roughness with a multi-scale wavelet analysis method [1]. This method allows quantifying each wavelength family composing the surfacewith the roughness criterion SMa, representing the average value of the signal amplitude for a given wavelength. The results are presented in the form of an SMa roughness criterion spectrum, ranging from small to large wavelengths, forming a unique and multispectral signature of the surface topography signal. Secondly, we measured the tactile sensorial vibrations of the samples using an electonical finger. Thirdlyl, we measured the tactile sensorial properties of the samples using an electonical finger and TouchFinger [2], an innovative device simulating an augmented human finger. When we touch the surface of an object with our index finger, the friction of the finger's skin on the surface vibrates the mechanoreceptors located in the finger pulp [3]. The vibratory response is then instantly converted into an electrical signal, transmitted via the central nervous system to the brain, which interprets this information as the sensory quality of the surface. Like the human finger, the device is equipped with vibration sensors and a force measurement sensor. It comes in the form of a half-ring that slips onto the tip of the index finger. When touching a surface, the vibrations and the force recorded by the device are transformed into tactile quality parameters for each of the four mechanoreceptors: Pacinian, Ruffini, Merkel, and Meissner [4][5]. The experimental protocol was conducted on a diversity of materials : cardboard, glass, aluminium, PET, PP, recycled PP, recycled PET; and a diversity of Textures achieved by mechanical or chemical treatment, spraying deposit and laser engraving . All samples were measured in topography with a white light confocal device, then characterized with the SMa parameter by the 2D continuous wavelet method. For the evaluation of the touch of the samples, a panel of 10 people aged 22 to 25 years was formed. Measurements were performed with the TouchFinger device in a temperature-controlled room (21°C +/- 2°C) and humidity (50% +/- 5%). For each surface, the panelists performed three successive touches and replicated the measurement three times. The vibratory signals obtained were analyzed by a frequency decomposition method to calculate the mean vibratory level (La parameter (dB)) in each of the mechanoreceptor frequency bands. The average values obtained of the La (dB) parameter for each of the mechanoreceptors were analyzed in correlation with the surface topography parameters. The results obtained allow establishing a database of the tactile quality of reference samples based on their material and texture.

The aim of this research is to detect re-entrant features which are a signature of the additive manufacturing process that involve sintering or melting of metal or plastic powders. Conventional surface measurement techniques such as profilometry or optical microscopy consider measured surface as a function z=z(x,y), meaning that for any given location on the surface (x, y), there is only one unique height (z). These techniques work well for surfaces whose shape can be locally approximated using a plane, cylinder, or ellipsoid, or for surfaces considered at a large scale of observation. For freeform surfaces with complex geometry, consisting of pores, voids, cracks, or overhangs being present at the microscale, and resulting from incompletely sintered or melted powder or the turbulent interaction of a laser energy beam, these so-called re-entrant features are impossible to register due to the physical limitations of optical and contact methods[1]. With the development of metrological high-resolution X-ray microcomputed tomography, it has become possible to study regions of surfaces invisible to other techniques. However, due to the implicit representation of the surface, a change in the approach to its parametric and non-parametric description, especially present in ISO or ASME standards, is necessary [2-3]. The subject of this study are parts produced using Selective Laser Melting technology. The input material Ti-6AL-4V powder with a grain size between 15 µm and 45 µm. The geometry of the samples was a quarter cylinder with a radius of 5 mm and a height of 25 mm. The samples were produced using EOS M290 machine. To achieve a large diversity of surface morphology, process parameters were varied: laser power (between 107 and 267 W) and scanning speed (between 250 mm/s and 1000 mm/s). The samples were not subjected to any additional finishing processing. The surface topography was registered using Baker Hughes/General Electric v|tome|x s240 microCT scanner, obtaining a voxel size of 5.5 µm. On each of the reconstructed 3D models of the sample surfaces, every 0.1 mm, 20 roughness profiles along the build-up direction were extracted. Each profile was averaged using a B-spline curve, assuming it as the local form. For each point of the original profile, its projection onto the form profile was determined, and then it was checked whether the coordinates of the subsequent projected points increased in the same direction as the coordinates of the form profile. If the local direction of increase changes, it means that a local re-entrant feature has been detected. A large number of such features may indicate inappropriately selected parameters of the technological process or disqualify other measurement techniques for credible registration of the surface topography. On the other hand, a small number of re-entrant features may suggest that an alternative measurement using profilometry or optical microscopy adequately reproduces the given SGP. The frequency of occurrence of these features, in the context of laser beam power and scanning speed, is the subject of analysis in this article. Exemplary results of detected re-entrant features and the effect of technological parameters on the frequency of their occurrence is shown in Figure 1.

Perfume is a complex composition of odorants and solvents, with a variety of physico-chemical characteristics such as volatility, molecular weight and vapor pressure, in addition to their unique olfactory properties. When perfume is applied to the skin, only part of its components is perceptible to the human sense of smell, and this perception depends on many factors such as the physico-chemical properties of each molecule, the interactions between the ingredients in the perfume formula and the surface to which it is applied, such as the skin. In addition, external factors can influence the evaporation of odorants, which can lead to significant variations in concentration profiles, even for similar doses. Given that the sensory performance of fragranced products is crucial, it is of great interest to understand how the skin can influence the behaviour of a fragrance. Consequently, methods are needed to characterize the interaction of fragrance with the human skin surface. Raman spectroscopy is proving to be a fast, non-invasive analytical technique, ideal for investigating the molecular mechanisms involved in the interactions between different types of fragrance and the human skin. In this study, we examined the performance of different fragrances on ex vivo skin as a function of time using Raman spectroscopy. We discuss the most important bands observed in the Raman spectra, and also present results concerning the characterization of different skin types, including proteins, sebum content and hydration level. Finally, we have classified these skin types according to these different parameters.

Deflectometry is a powerful measuring technique of complex optical surfaces. It has extremely high sensitivity to surface slopes, but the in most cases, the measurement accuracy can only achieve a level of microns. The main limiting factors of the measurement accuracy are investigated. A flexible camera model and an automatic calibration method are developed. The geometrical pose and location of the workpiece are solved by minimizing the re-projection error of ray-tracing, and an automatic positioning method is proposed by combining the Gaussian process regression.

Most predatory theropod dinosaurs, excluding spinosaurs, have long been thought to retain the plesiomorphic dinosaurian condition of carnivory. The presence of numerous, similar sympatric medium- to large-bodied predatory taxa in localities such as the Bahariya, Huincul and Morrison formations therefore raise questions regarding ecosystem structure. Spinosaurs are known for piscivory and possess adaptations for subaqueous feeding and locomotion, whereas their sister clade, megalosaurids have generally been considered to include only terrestrial predators. However, increasing evidence suggests that megalosaurids, especially European taxa, had a preference for feeding in mesic and aquatic environments. Establishing dinosaurian diets have often relied on tooth morphology or stomach contents, both which have limitations, such as convergent morphology in the case of the former or capturing fall-back foods rather than usual diet in the case of the latter. Moreover, tooth morphology usually recognises broad dietary guilds, such as herbivory or carnivory, only. Recent work on extant and extinct archosaurs and lepidosaurs have shown that dietary guilds can be determined for fossil taxa more precisely, including categories like hard or soft invertebrate diets, carnivory and piscivory, among others, by using Dental Microwear Texture Analysis (DMTA)1,2,3. Using a confocal microscopy and dental moulds, the 3D surface textures of a theropod tooth sample were analysed at 100x magnification using common parameters from the literature 1,2 . Preliminary results suggest that predatory theropod diets were more diverse than previously thought, with megalosaurids showing evidence for piscivory and tyrannosauroids (the clade that includes Tyrannosaurus rex) potentially exhibiting non-carnivory in their diets. Consequently, niche partitioning might have occurred between sympatric theropod taxa due to specialisations in their diets that prevented direct competition. Furthermore, application of DMTA along the tooth rows within individual theropod jaws might allow the method of prey acquisition to be determined based on serial variation in the microwear patterns of the teeth with the potential to test biomechanical models.