This PDF file contains the front matter associated with SPIE Proceedings Volume 11732, including the Title Page, Copyright Information, and Table of Contents.

Introduction to SPIE Defense and Commercial Sensing conference 11732: Dimensional Optical Metrology and Inspection for Practical Applications X

3D imaging sensors are important for intelligent robots to figure out complicated and unknown assembly tasks. To perform a smooth robotic assembly with high precision, it is crucial to incorporate 3D imaging sensors that can scan, recognize, and characterize fine-scale features (e.g., small screws). We developed an integrated robotic 3D vision system based on high-resolution structured light 3D vision based on fringe projection. Our preliminary testing has demonstrated the promise of our integrated system in precision assembly operations supported by the capability of extracting fine-scale key features through high-resolution surface mapping.

Structured light sensors based on diffuse reflection of projected patterns are widely used to measure the 3D shape of objects, e.g., in industry, medicine, or cultural heritage. Unfortunately, there exist many objects made of uncooperative materials, i.e., materials with optical properties such as being glossy, transparent, absorbent, or translucent, which cannot be measured reliably by this measurement technique. In the last years, we presented a two-step method based on thermal pattern projection which allows the determination of the object surface of these uncooperative objects. In the first step, a multi-fringe thermal pattern is projected onto the measurement object. In the second step, a mid-wave infrared (MWIR) stereo camera records the thermal radiation that is absorbed and re-emitted by the object surface. This system allows us to measure transparent objects. However, the measurement time is with tens of seconds up to minutes quite long and the measurement accuracy should be improved. In this contribution, we present a projection principle of a fast sequentially scanning fringe leading to a significant reduction in measurement periods down to the second range or even below while increasing the measurement accuracy. The work includes a characterization of our MWIR 3D setup for both projection principles, the multifringe and sequential fringe one, regarding the measurement accuracy and speed. We show measurement results obtainted with both projection techniques for an object made of different material classes.

Pattern projection-based stereo 3D sensors are widely used for contactless, non-destructive optical 3D shape measurements. In previous works, we have shown 3D measurement systems based on stereo matching between two cameras with GOBO-projected aperiodic fringe patterns. We have also demonstrated a low latency 3D reconstruction algorithm (BICOS), which can be used for real time 3D measurements. We showed an optimization method for the projected aperiodic fringe patterns with the purpose of making the measurements more robust and to reduce the pattern sequence length without sacrificing 3D model completeness. In this contribution, we demonstrate a sensor for a medical application which aggregates these developments. Our sensor is used to monitor patient movement during radiation therapy. In this application a low measurement latency is of high importance. A significant part of this latency is caused by image acquisition. We show that we can reduce the number of required image pairs to 6 when optimizing the projected aperiodic fringe patterns. In combination with our BICOS algorithm, we can achieve total measurement latencies of below 80 ms at an accuracy of 355 μm in a measurement field of 1 m × 2 m.

There are many methods used today to capture general 3D scene information including laser radar, structured light and stereo imaging. For a wide, general scene, laser radar and structured light projection requires active illumination which for a wide, outdoor view may be impractical, either because of the high-power light needed, the desired to be discrete, or just to time needed to scan a wide scene using a beam or pattern that does not blind persons in the field. Stereo can use passive light, but requires a larger stereo capture separation (multiple camera separation) and good correlation points that may be had to find. This paper will present the use of passive IR imaging along with focus diversity to capture a wide scene with just a few image fields. Similar work has been reported for capturing fingerprints and hand-prints at high resolution as well as face information at large distances. The paper will review the background and theory and present some illustrative results.

Fringe projection profilometry (FPP) is a non-contact, high-precision technique for measuring three-dimensional (3D) shapes. An essential step of FPP is to recover the phase distribution from the deformed fringe patterns. In real applications, the captured fringe patterns often suffer from noises, which results in degradation of the performance of phase retrieval and shape reconstruction. Fringe denoising can be applied to suppress the influence of noise in FPP. This paper introduces a novel fringe denoising method based on robust principal component analysis (RPCA). The proposed method makes use of the low-rankness of the clean fringe patterns and the sparsity of the strong impulsive fringe noise. RPCA is then applied to effectively mitigate the strong impulsive fringe noise and suppress the random additive noise. The proposed method features 2D processing of the fringe patterns and is easy to implement. Its effectiveness is demonstrated via numerical simulations.

Phase-shifting profilometry (PSP) is one of the mainstream fringe projection techniques for object surface reconstruction. Thanks to its multi-shot nature, PSP is less sensitive to ambient light and reflectivity on the object surface. For static objects, the performance of PSP can be improved by projecting and capturing more phase-shifted fringe patterns. However, when applied to dynamic objects, PSP suffers from motion-induced errors due to the loss of correspondence. This paper proposes a new approach to improving the performance of measuring objects with a general three-dimensional (3D) movement. Firstly, instead of employing a large number of fringe patterns for PSP, we apply three-step PSP, which suffers less from motion errors, to obtain multiple coarse measurements of the object. Then the moving object is segmented, and the iterative closest point (ICP) algorithm is applied to estimate the motion parameters. Finally, the multiple measurements of the object are fused using adaptive weights. The proposed scheme alleviates the motion errors of PSP with a large number of fringe patterns and enhances the accuracy of three-step PSP. Simulations and experiments verify the feasibility of the proposed scheme.

Industrial parts often have surface finish requirements as part of their product acceptance criteria. In this study, the authors will review and present data studying the performance capabilities of the Chromatic Line Confocal Sensing (LCI) for quantifying part surface finish. Accuracy over a range of surface finish conditions will be studied and presented. Further, a comparison of LCI vs. tactile measurements results will be explored and discussed.

Within the paper the use of a 3D-line scanner is evaluated to realize an inline and intime detection of weld imperfections for an automated MAG-welding of fillet welds. The required conditions of the weld are in accordance with the quality level B of the international standard ISO 5817 [1]. The used system is a Wenglor 3D-line scanner with a wavelength of 660 nm and an image recording frequency of at least 200 Hz. This scanner is able to detect, as a stand-alone system, geometry and surface imperfections of the weld. Therefore, algorithms were developed for the reliable recognition of deviations regarding the size, superelevations, undercuts, seam transition angles and the symmetry of the weld with a measurement accuracy of 10 ➭m. With the information, gathered by the line scanner and the corresponding image-data processing algorithms, it is possible to improve the efficiency of automated MAG-welding. The welding process can be interrupted and recalibrated if significant imperfections are detected. This is realized with a short time delay depending on the distance between the welding torch and the camera and the welding speed. The direct feedback could prevent the weldment from getting unusable and saving ressources through minor required efforts for renewing or revising the weld. Additionally, the required visual testing (ISO 17635 [2], ISO 17637 [3]) can be objectified and accelerated by highlighting the parts of the weld which are most probably affected by imperfections. In the future this automated system could possibly replace the manual quality inspection.

Toroidal end mills have great industrial applications and have good prospects for increasing the volume of products produced with them. It complicates the process of measuring the cutting edges by changing the position of the cutting edges along the axis, which does not allow the geometry control at the same focal length. The method proposed in the article is an effective tool for carrying out automated measurements of the shapes of cutting edges. Application of the method is relevant for edge control in automated mode. To assess the accuracy of the proposed method, it was tested on a high precision measuring machine Walter Helicheck Plus. The new method allows for repeatability of measurements down to 1.5 μm, which meets the accuracy requirements for non-profiling areas of end mills. Suggested method is universal and can significantly reduce the laboriousness of measuring multi-flute cutters. The most important advantage of this method is the ability to determine the ranges in which there is a smooth cutting edge in all areas of the cutter's cutting edge.

A novel original method is presented to detect and track laser damage sites on vacuum windows of the Laser MegaJoule (LMJ) facility. The method is based on spatial registration by Digital Image Correlation (DIC). It also involves corrections for gray level variations induced by variable lighting conditions. Using the present method, an efficient way is achieved to detect and follow laser damage sites as soon as they appear on the optical component. The developed tools offer the possibility of characterizing and predicting damage growth as a function of laser shot features.

The article developed an approach to automated analysis images obtained in the visible and thermal imaging range by a system consisting of two separate cameras. For the generated images, a technique is proposed for the parallel application of algorithms for primary data processing (filtering and blur restoration method), searching for communication elements between data and performing geometric transformations over them, necessary for their combination. To identify elements with elevated temperatures, a primary (normal) thermal gradient mask is created. For simplicity, the resulting mask image is simplified. The simplification method is based on a nonlinear change in the image's bitness by analyzing histograms of color gradients. Next, a mask of threshold values for objects recorded by the optical system is formed. The paper proposes an algorithm for the averaged comparison of the received data and identifying areas where the values are exceeded. The advantage of the developed approach is its versatility. The use of one tuned system makes it possible to signal the cutting tool's wear, the presence of inclusions (inhomogeneities) in the workpiece part, wear of the bearings or elements of the robotic complex (machine tool). The set of test data provides an example of applying the proposed approach for analyzing the production process of cutting metal using a waterjet machine and milling a part on a numerically controlled machine.

In the modern world, the use of visual information is ubiquitous. Small-sized data analysis devices have found application in almost all areas, including security systems, industrial production, gaming, car control, etc. To increase the accuracy, fault tolerance, and decision-making speed, combined information data fields are used as input data. The article developed an approach to combining data obtained in different ranges into a single image. To improve the alignment accuracy, the paper proposes using a group of methods for the primary processing of image enhancement. The article proposes finding anchor points for a series of images used for their subsequent merging. An algorithm for calibrating a system of several cameras is presented, a description of the influence of parameters on the result of combining is given. To increase the processing speed and the selection of the processing methods' primary parameters, it is proposed to change the bitness of the image by simplifying them. The operation includes absorbing small objects and averaging the ranges of histograms of color gradients. As test data, a series of images of one object obtained in: visible (camera with a resolution of 1920 x 1080 pixels, color depth of 8 bits); far-infrared (thermal imaging images with a resolution of 320x240 pixels, grayscale); 3D model (construction is carried out based on a stereo pair with a resolution of 1024x768 pixels).