The application of digital geometry processing is undergoing an extension from small industrial products to large-scale structures and environments, including plants, factories, ships, bridges, buildings, forests, and indoor/outdoor/urban environments. This extension is being supported by recent advances in long-range 3D laser scanning technology. Laser scanners are mounted on various platforms, such as tripods, wheeled vehicles, airplanes, and UAVs, and the laser scanning systems are used to efficiently acquire dense and accurate digitized 3D data of the geometry, called point clouds, of large-scale structures and environments. As another technology for the acquisition of digital 3D data of structures and environments, 3D reconstruction methods from digital images are also attracting a great deal of attention because of their flexibility.
The utilization of digital 3D data for various purposes still has many challenges, however, in terms of data processing. The extraction of accurate and meaningful information from the data is an especially important and difficult problem, and many studies on object and scene recognition are being conducted in many fields. How to acquire useful and high-quality digital 3D data of large-scale structures and environments is another problem to be solved for digital geometry processing to be widely used.
This special issue addresses the latest research advances in digital geometry processing for large-scale structures and environments. It covers a broad range of topics in geometry processing, including new technologies, systems, and reviews for 3D data acquisition, recognition, and modeling of ships, factories, plants, forests, river dikes, and urban environments.
The papers will help the readers explore and share their knowledge and experience in technologies and development techniques in this area. All papers were refereed through careful peer reviews. We would like to express our sincere appreciation to the authors for their excellent submissions and to the reviewers for their invaluable efforts in producing this special issue.
Mobile mapping systems can capture point clouds and digital images of roadside objects. Such data are useful for maintenance, asset management, and 3D map creation. In this paper, we discuss methods for extracting guardrails that separate roadways and walkways. Since there are various shape patterns for guardrails in Japan, flexible methods are required for extracting them. We propose a new extraction method based on point processing and a convolutional neural network (CNN). In our method, point clouds and images are segmented into small fragments, and their features are extracted using CNNs for images and point clouds. Then, features from images and point clouds are combined and investigated using whether they are guardrails or not. Based on our experiments, our method could extract guardrails from point clouds with a high success rate.
Herbaceous vegetation on riverdikes plays an important role in preventing soil erosion, which, otherwise, may lead to the collapse of riverdikes and consequently, severe flooding. It is crucial for managers to keep suitable vegetation conditions, which include native grass species such as Imperata cylindrica, and to secure visibility of riverdikes for inspection. If managers can efficiently find where suitable grass and unsuitable forb species grow on vast riverdikes, it would help in vegetation management on riverdikes. Classification and quantification of herbaceous vegetation is a challenging task. It requires spatial resolution and accuracy high enough to recognize small, complex-shaped vegetation on riverdikes. Recent developments in unmanned aerial vehicle (UAV) technology combined with light detection and ranging (LiDAR) may offer the solution, since it can provide highly accurate, high-spatial resolution, and denser data than conventional systems. This paper aims to develop a model to classify grass and forb species using UAV LiDAR data alone. A combination of UAV LiDAR-based structural indices, V-bottom (presence of vegetation up to 50 cm from the ground) and V-middle (presence of vegetation 50–100 cm from the ground), was tested and validated in 94 plots owing to its ability to classify grass and forb species on riverdikes. The proposed method successfully classified the “upright” grass species and “falling” grass species / forb species with an accuracy of approximately 83%. Managers can efficiently prioritize the inspection areas on the riverdikes by using this method. The method is versatile and adjustable in other grassland environments.
Laser measurement technology has progressed significantly in recent years, and diverse methods have been developed to measure three-dimensional (3D) objects within environmental spaces in the form of point cloud data. Although such point cloud data are expected to be used in a variety of applications, such data do not possess information on the specific features represented by the points, making it necessary to manually select the target features. Therefore, the identification of road features is essential for the efficient management of point cloud data. As a technology for identifying features from the point cloud data of road spaces, in this research, we propose a method for automatically dividing point cloud data into units of features and identifying features from projected images with added depth information. We experimentally verified that the proposed method accurately identifies and extracts such features.
In this study, we developed a new system that outputs the additional press work procedures necessary to obtain the desired ship-hull surface. This study is unique in terms of determining the additional press work procedures required according to the current plate shape at any work stage by measuring the plate shape using a laser scanner. In the proposed method, a B-spline surface is constructed from a point cloud measured using a laser scanner, and the current plate shape is analyzed based on differential geometry. Additional press lines are estimated based on the difference in the normal curvature along the lines of curvature between the designed target surface and the current surface. We demonstrated the effectiveness of our proposed method through experiments at a shipyard. The proposed system may be used to enhance the efficiency of press work and is expected to be an effective tool for training beginners in the future.
Digital image phenotyping has become popular in plant research. Plants are complex in shape, and occlusion can often occur. Three-dimensional (3D) data are expected to measure the morphological traits of plants with higher accuracy. Plants have organs with flat and/or narrow shapes and similar component structures are repeated. Therefore, it is difficult to construct an accurate 3D model by applying methods developed for industrial materials and architecture. Here, we review noncontact and all-around 3D modeling and configuration of camera systems to measure the morphological traits of plants in terms of system composition, accuracy, cost, and usability. Typical noncontact 3D measurement methods can be roughly classified into active and passive methods. We describe their advantages and disadvantages. Structure-from-motion/multi-view stereo (SfM/MVS), a passive method, is the most frequently used measurement method for plants. It is described in terms of “forward intersection” and “backward resection.” We recently developed a novel SfM/MVS approach by mixing the forward and backward methods, and we provide a brief overview of our approach in this paper. While various fields are adopting 3D model construction, nonexpert users struggle to use them and end up selecting inadequate methods, which lead to model failure. We hope that this review will help users who are considering starting to construct and measure 3D models.
In this study, we develop a system for efficiently measuring detailed information of trees in a forest environment using a small unmanned aerial vehicle (UAV) equipped with light detection and ranging (lidar). The main purpose of forest measurement is to predict the volume of wood for harvesting and delineating forest boundaries by tree location. Herein, we propose a method for extracting the position, number of trees, and vertical height of trees from a set of three-dimensional (3D) point clouds acquired by a UAV lidar system. The point cloud obtained from a UAV is dense in the tree’s crown, and the trunk 3D points are sparse because the crown of the tree obstructs the laser beam. Therefore, it is difficult to extract single-tree information from 3D point clouds because the characteristics of 3D point clouds differ significantly from those of conventional 3D point clouds using ground-based laser scanners. In this study, we segment the forest point cloud into three regions with different densities of point clouds, i.e., canopy, trunk, and ground, and process each region individually to extract the target information. By comparing a ground laser survey and the proposed method in an actual forest environment, it is discovered that the number of trees in an area measuring 100 m × 100 m is 94.6% of the total number of trees. The root mean square error of the tree position is 0.3 m, whereas that of the vertical height is 2.3 m, indicating that single-tree information can be measured with sufficient accuracy for forest management.
Recently, three-dimensional (3D) laser scanning technology using terrestrial laser scanner (TLS) has been widely used in the fields of plant manufacturing, civil engineering and construction, and surveying. It is desirable for the operator to be able to immediately and intuitively confirm the scanned point cloud to reduce unscanned regions and acquire scanned point clouds of high quality. Therefore, in this study, we developed a method to superimpose the point cloud on the actual environment to assist environmental 3D laser measurements, allowing the operator to check the scanned point cloud or unscanned regions in real time using the camera image. The method included extracting the correspondences of the camera image and the image generated by point clouds by considering unscanned regions, estimating the camera position and attitude in the point cloud by sampling correspondence points, and superimposing the scanned point cloud and unscanned regions on the camera image. When the proposed method was applied to two types of environments, that is, a boiler room and university office, the estimated camera image had a mean position error of approximately 150 mm and mean attitude error of approximately 1°, while the scanned point cloud and unscanned regions were superimposed on the camera image on a tablet PC at a rate of approximately 1 fps.
Monitoring technologies have attracted attention in the factory automation fields that rely on the Internet of Things (IoT). However, it is difficult to monitor the process information from a round machining tool during rotating operations. Therefore, we developed a novel tool holder equipped with a wireless communication function to monitor tool vibrations. In the present study, we attempt to measure the tool holder vibrations during ball nose end milling processes using the servo driving information for different machine tools. We demonstrate that, using the developed tool holder with a wireless system, it is feasible to improve the machined free form surface by considering the servo driving information.
The aim of this study is to show the effect of the strain-rate on the forming limit strain of an aluminum alloy A5052 sheet and a mild steel sheet SPCC. Biaxial stretching test was carried out. The prescribed strain path was linear path or that with directional change in straining. The sheet was pre-strained by uniaxial tension in the latter path. The deformation speed was set to be quasi-static or high speed whose strain-rate was about 300 /s using the dedicated high speed stretching device. The forming limit strain of the A5052 sheet for the linear strain path was larger in the high speed stretching than that under the quasi-static condition. For the case with strain path change the forming limit strain was further large. This may be due to the softening phenomenon which occurs by aging treatment, because the stretching experiment was conducted about two weeks after the pre-straining operation. On the other hand, the forming limit strain of the SPCC under the high speed condition was smaller than that under the quasi-static condition in the linear strain path. This is attributed to the decreased strain hardening exponent when the strain-rate increases. Further, in the equi-biaxial stretching of the pre-strained specimen, large difference of the forming limit strain between the deformation speeds was found. It is concluded that A5052 aluminum alloy sheet has a good adaptability to high speed forming, on the other hand, attention should be paid in increasing the forming speed of SPCC.
Machining time estimation is essential for the due-date estimation of products as well as for production planning. Conventionally, machining time has been estimated by a computer aided manufacturing (CAM) system, which requires time and effort to create its numerical control (NC) program and requires machining expertise to operate it. In addition, among the problems with conventional methods, an error in the estimated machining time arises owing to the machine tool’s control characteristics. In this study, an artificial intelligence (AI)-based system capable of estimating machining time promptly and simply based on shape data without requiring any NC program is developed. The input data to the AI system are color information regarding the machined depths, which are used to estimate the rough-machining time, and color information regarding the machined surface curvature distributions to estimate the finish-machining time. Color information on the machined depths and machined surface curvature distributions is created using three-dimensional computer aided design (3D CAD) data. To build the AI system, the shape data and machining time data accumulated at the machining site are used, so that the machining time estimated reflects the machining method, machining expertise, and the machine tool characteristics employed.
This study investigated a method for accurately predicting the residual stress in die castings manufactured using aluminum alloy. To account for the mechanical properties caused by the material composition differences that occur in the thickness direction of the die castings, a model split in the thickness direction was used in the simulation model. Norton’s law was applied to the constitutive equation of the material, and the stress relaxation phenomenon was examined. The composition of Al-Si-Cu alloy (JIS-ADC12) die castings in the thickness direction were analyzed using scanning electron microscopy and energy dispersive X-ray spectroscopy (SEM-EDS), and differences in composition were confirmed. As a result of calculating the residual stress using the simulation, it was possible to calculate the residual stress that could not be reproduced by the simulation model of uniform composition. This suggested that the difference in mechanical properties of die castings in the micro-region influences the residual stress.
A novel plastic injection mold design for a product with a deep arc-hole based on the kinematic analysis is proposed. To move the wide angle arc-core, a hybrid mechanism combining a slider-crank and a swing cam is designed. The force transmission coefficient of the slider-crank and the pressure angle of the swing cam are used to evaluate the dimensions of the mechanism. The motion of the designed mechanism was confirmed by using a prototype made with a 3D printer.