Non-destructive analysis of flake properties in automotive paints with full-field optical coherence tomography and 3D segmentation

A Review of In Situ Defect Detection and Monitoring Technologies in Selective Laser Melting

The additive manufacturing (AM) technique has received considerable industrial attention, as it is capable of producing complex functional parts in the aerospace and defense industry. Selective laser melting (SLM) technology is a relatively mature AM process that can manufacture complex structures both directly and efficiently. However, the quality of SLM parts is affected by many factors, resulting in a lack of repeatability and stability of this method. Therefore, several common and advanced in situ monitoring as well as defect detection methods are utilized to improve the quality and stability of SLM processes. This article aims at documenting the various defects that occurred in SLM processes and their influences on the final parts. Various types of in situ monitoring and defect detection methods and their applications are reviewed, and their integrations with the SLM processes are also discussed.

Multiclass Level-Set Segmentation of Rust and Coating Damages in Images of Metal Structures

This paper describes the combined detection of coating and rust damages on painted metal structures through the multiclass image segmentation technique. Our prior works were focused solely on the localization of rust damages and rust segmentation under different ambient conditions (different lighting conditions, presence of shadows, low background/object color contrast). This paper method proposes three types of damages: coating crack, coating flaking, and rust damage. Background, paint flaking, and rust damage are objects that can be separated in RGB color-space alone. For their preliminary classification SVM is used. As for paint cracks, color features are insufficient for separating it from other defect types as they overlap with the other three classes in RGB color space. For preliminary paint crack segmentation we use the valley detection approach, which analyses the shape of defects. A multiclass level-set approach with a developed penalty term is used as a framework for the advanced final damage segmentation stage. Model training and accuracy assessment are fulfilled on the created dataset, which contains input images of corresponding defects with respective ground truth data provided by the expert. A quantitative analysis of the accuracy of the proposed approach is provided. The efficiency of the approach is demonstrated on authentic images of coated surfaces.

Methods and applications of full-field optical coherence tomography: a review

SIGNIFICANCE

Full-field optical coherence tomography (FF-OCT) enables en face views of scattering samples at a given depth with subcellular resolution, similar to biopsy without the need of sample slicing or other complex preparation. This noninvasive, high-resolution, three-dimensional (3D) imaging method has the potential to become a powerful tool in biomedical research, clinical applications, and other microscopic detection.

AIM

Our review provides an overview of the disruptive innovations and key technologies to further improve FF-OCT performance, promoting FF-OCT technology in biomedical and other application scenarios.

APPROACH

A comprehensive review of state-of-the-art accomplishments in OCT has been performed. Methods to improve performance of FF-OCT systems are reviewed, including advanced phase-shift approaches for imaging speed improvement, methods of denoising, artifact reduction, and aberration correction for imaging quality optimization, innovations for imaging flux expansion (field-of-view enlargement and imaging-depth-limit extension), new implementations for multimodality systems, and deep learning enhanced FF-OCT for information mining, etc. Finally, we summarize the application status and prospects of FF-OCT in the fields of biomedicine, materials science, security, and identification.

RESULTS

The most worth-expecting FF-OCT innovations include combining the technique of spatial modulation of optical field and computational optical imaging technology to obtain greater penetration depth, as well as exploiting endogenous contrast for functional imaging, e.g., dynamic FF-OCT, which enables noninvasive visualization of tissue dynamic properties or intracellular motility. Different dynamic imaging algorithms are compared using the same OCT data of the colorectal cancer organoid, which helps to understand the disadvantages and advantages of each. In addition, deep learning enhanced FF-OCT provides more valuable characteristic information, which is of great significance for auxiliary diagnosis and organoid detection.

CONCLUSIONS

FF-OCT has not been completely exploited and has substantial growth potential. By elaborating the key technologies, performance optimization methods, and application status of FF-OCT, we expect to accelerate the development of FF-OCT in both academic and industry fields. This renewed perspective on FF-OCT may also serve as a road map for future development of invasive 3D super-resolution imaging techniques to solve the problems of microscopic visualization detection.

Non-Destructive Subsurface Inspection of Marine and Protective Coatings Using Near- and Mid-Infrared Optical Coherence Tomography

Near- and mid-infrared optical coherence tomography (OCT) is evaluated as a non-destructive and non-contact reflection imaging modality for inspection of industrial and marine coatings. Near-infrared OCT was used to obtain high-resolution images (~6/2 µm lateral/axial) of hidden subsurface cracks and defects in a resin base coating, which had been exposed to high pressure and high temperature to study coating degradation in hostile environments. Mid-infrared OCT was employed for high-resolution (~15/8.5 µm lateral/axial) subsurface inspection of highly scattering marine coatings, demonstrating monitoring of wet film thickness and particle dispersion during curing of a 210 µm layer of antifouling coating, and detection of substrate corrosion through 369 µm of high-gloss alkyd enamel. Combining high-resolution and fast, non-invasive scanning, OCT is therefore considered a promising tool for studying coating performance and for industrial inspection.

Deep learning architecture "LightOCT" for diagnostic decision support using optical coherence tomography images of biological samples

Optical coherence tomography (OCT) is being increasingly adopted as a label-free and non-invasive technique for biomedical applications such as cancer and ocular disease diagnosis. Diagnostic information for these tissues is manifest in textural and geometric features of the OCT images, which are used by human expertise to interpret and triage. However, it suffers delays due to the long process of the conventional diagnostic procedure and shortage of human expertise. Here, a custom deep learning architecture, LightOCT, is proposed for the classification of OCT images into diagnostically relevant classes. LightOCT is a convolutional neural network with only two convolutional layers and a fully connected layer, but it is shown to provide excellent training and test results for diverse OCT image datasets. We show that LightOCT provides 98.9% accuracy in classifying 44 normal and 44 malignant (invasive ductal carcinoma) breast tissue volumetric OCT images. Also, >96% accuracy in classifying public datasets of ocular OCT images as normal, age-related macular degeneration and diabetic macular edema. Additionally, we show ∼96% test accuracy for classifying retinal images as belonging to choroidal neovascularization, diabetic macular edema, drusen, and normal samples on a large public dataset of more than 100,000 images. The performance of the architecture is compared with transfer learning based deep neural networks. Through this, we show that LightOCT can provide significant diagnostic support for a variety of OCT images with sufficient training and minimal hyper-parameter tuning. The trained LightOCT networks for the three-classification problem will be released online to support transfer learning on other datasets.

Polarization-coded material classification in automotive LIDAR aiming at safer autonomous driving implementations

LIDAR sensors are one of the key enabling technologies for the wide acceptance of autonomous driving implementations. Target identification is a requisite in image processing, informing decision making in complex scenarios. The polarization from the backscattered signal provides an unambiguous signature for common metallic car paints and can serve as one-point measurement for target classification. This provides additional redundant information for sensor fusion and greatly alleviates hardware requirements for intensive morphological image processing. Industry decision makers should consider polarization-coded LIDAR implementations. Governmental policy makers should consider maximizing the potential for polarization-coded material classification by enforcing appropriate regulatory legislation. Both initiatives will contribute to faster (safer, cheaper, and more widely available) advanced driver-assistance systems and autonomous functions. Polarization-coded material classification in automotive applications stems from the characteristic signature of the source of LIDAR backscattering: specular components preserve the degree of polarization while diffuse contributions are predominantly depolarizing.

Sub-surface imaging of soiled cotton fabric using full-field optical coherence tomography

In the laundry industry, colorimetry is a common way to evaluate the stain removal efficiency of detergents and cleaning products. For ease of visualization, the soiling agent is treated with a dye before measurement. However, it effectively measures the dye removal rather than stain removal, and it cannot provide depth-resolved information of the sample. In this study, we show that full-field (FF) optical coherence tomography (OCT) technique is capable of measuring the cleaning effect on cotton fabric by imaging the sub-surface features of fabric samples. We used a broadband light-emitting diode (LED) source to power the FF-OCT system that achieves the resolution of 1 µm axially and 1.6 µm laterally. This allows the micron-sized cotton fibres/fibrils at different depth positions to be resolved. The clean, the soiled, and the washed samples can be differentiated from their cross-sectional images using OCT, where the cleaning effect can be correlated with the sub-surface fibre volume. The experimental results of the proposed method were found to be in good agreement with those of the standard colorimetry method. The proposed technique therefore offers an alternative way for measuring the stain removal from fabric substrate to assess the effectiveness of laundry detergent products.

Roles of Extracellular Vesicles in High-Grade Gliomas: Tiny Particles with Outsized Influence

High-grade gliomas, particularly glioblastomas (grade IV), are devastating diseases with dismal prognoses; afflicted patients seldom live longer than 15 months, and their quality of life suffers immensely. Our current standard-of-care therapy has remained essentially unchanged for almost 15 years, with little new therapeutic progress. We desperately need a better biologic understanding of these complicated tumors in a complicated organ. One area of rejuvenated study relates to extracellular vesicles (EVs)-membrane-enclosed nano- or microsized particles that originate from the endosomal system or are shed from the plasma membrane. EVs contribute to tumor heterogeneity (including the maintenance of glioma stem cells or their differentiation), the impacts of hypoxia (angiogenesis and coagulopathies), interactions amid the tumor microenvironment (concerning the survival of astrocytes, neurons, endothelial cells, blood vessels, the blood-brain barrier, and the ensuing inflammation), and influences on the immune system (both stimulatory and suppressive). This article reviews glioma EVs and the ways that EVs manifest themselves as autocrine, paracrine, and endocrine factors in proximal and distal intra- and intercellular communications. The reader should note that there is much controversy, and indeed confusion, in the field over the exact roles for EVs in many biological processes, and we will engage some of these difficulties herein.

Fast esophageal layer segmentation in OCT images of guinea pigs based on sparse Bayesian classification and graph search

Endoscopic optical coherence tomography (OCT) devices are capable of generating high-resolution images of esophageal structures at high speed. To make the obtained data easy to interpret and reveal the clinical significance, an automatic segmentation algorithm is needed. This work proposes a fast algorithm combining sparse Bayesian learning and graph search (termed as SBGS) to automatically identify six layer boundaries on esophageal OCT images. The SBGS first extracts features, including multi-scale gradients, averages and Gabor wavelet coefficients, to train the sparse Bayesian classifier, which is used to generate probability maps indicating boundary positions. Given these probability maps, the graph search method is employed to create the final continuous smooth boundaries. The segmentation performance of the proposed SBGS algorithm was verified by esophageal OCT images from healthy guinea pigs and the eosinophilic esophagitis (EoE) models. Experiments confirmed that the SBGS method is able to implement robust esophageal segmentation for all the tested cases. In addition, benefiting from the sparse model of SBGS, the segmentation efficiency is significantly improved compared to other widely used techniques.

High-performance optical coherence velocimeter: theory and applications

We proposed a high-performance optical coherence velocimeter (OCV) based on broadband optical interference which achieves spatial resolution from interference cancellation or enhancement of different components of the broadband light. There is a challengeable issue for OCV that the interference fringes become blurred when the velocity of detected object is relatively large, hindering the pace of OCV application in high-velocity field. To resolve this, the relationship between blurry coefficient and OCV system parameters (e.g., exposure time, central wavelength, bandwidth of source) was derived. It was found that blurry coefficient changed with oscillatory decay form and reached the minimum at each order blurry velocity. It showed that maximum measurable velocity of OCV systems could reach 10th order blurry velocity. The measurement of vibration of the loudspeaker driven by a function signal generator was employed to experimentally verify the velocity measurement performance of the system. The experiment demonstrated that the developed OCV can provide large velocity measurement ranges from static to 25.2 mm/s with nanometer-level precision and maximum measurable vibration frequency of up to 50 kHz. However, in theory, the theoretical maximum measurable velocity can be up to 1.06 m/s for current OCV configuration. The OCV has high precision, large dynamic range, and high-velocity measurement capability, making it attractive for applications in mechanical structure vibration monitoring and acoustic measurement.

In situ process monitoring in selective laser sintering using optical coherence tomography

Selective laser sintering (SLS) is an efficient process in additive manufacturing that enables rapid part production from computer-based designs. However, SLS is limited by its notable lack of in-situ process monitoring when compared to other manufacturing processes. We report the incorporation of optical coherence tomography into an SLS system in detail and demonstrate access to surface and sub-surface features. Video frame rate cross-sectional imaging reveals areas of sintering uniformity and areas of excessive heat error with high temporal resolution. We propose a set of image processing techniques for SLS process monitoring with OCT and report the limitations and obstacles for further OCT integration with SLS systems.

High-resolution retinal swept source optical coherence tomography with an ultra-wideband Fourier-domain mode-locked laser at MHz A-scan rates

We present a new 1060 nm Fourier domain mode locked laser (FDML laser) with a record 143 nm sweep bandwidth at 2∙ 417 kHz  =  834 kHz and 120 nm at 1.67 MHz, respectively. We show that not only the bandwidth alone, but also the shape of the spectrum is critical for the resulting axial resolution, because of the specific wavelength-dependent absorption of the vitreous. The theoretical limit of our setup lies at 5.9 µm axial resolution. In vivo MHz-OCT imaging of human retina is performed and the image quality is compared to the previous results acquired with 70 nm sweep range, as well as to existing spectral domain OCT data with 2.1 µm axial resolution from literature. We identify benefits of the higher resolution, for example the improved visualization of small blood vessels in the retina besides several others.