A method to utilize three-dimensional point cloud data obtained by using mobile mapping system (MMS) for landscape management of cities is necessary, but still undeveloped. In this paper, we focus on spatial distribution features of reflection intensities included in point cloud data and investigated whether an index on spatial autocorrelation is useful or not for determing the presence of leaves in a tree. In experiments, we modeled the three-dimensional space in the voxel space and calculated the indicators of spatial autocorrelation (Moran's I statistic, Lee's L statistic and indicators decomposed from these) with the mean and (or) standard deviation of reflection intensities for each voxel. As a result of the experiments, Lee's L statistic and indicators decomposed from this, that were calculated with the mean and standard deviation of reflection intensities for each voxel, can be useful for determing the presence of leaves in a tree.

View full abstract

 CIM and i-Construction promoted by the Ministry of Land, Infrastructure and Transport and Tourism are premised on 3D models. However, annotations of 3D models, there is no standard in the civil engineering field and they depend on a commercial 3D CAD software. In this research, we conducted a survey on 3D annotated models for the automobile industry where the use of 3D information technology is in advanced. As a result, it was found that the automobile industry has created 3D annotated models in compliance with ISO 16792. Therefore, when we investigate ISO 16792 and create 3D annotated models for civil engineering structures, it was possible to extract matters to be complied with anotinations and evaluate implementation and verification of annotations.

View full abstract

 Civil struture drawing is discribed design intents and tolerance is stated in the measurement items and the standard values described in work progress control stadard. Meanwhile, drawings of the manufacturing industries, along with design values and tolerance are also described using dimensional tolerances.
 In this research, we decided to apply dimensional tolerance of bridge piers, steel pipe soil cement piles and rebar of piers. As a result, it was possible to apply dimensional tolerances that conform to the measurement items and specifications described in work progress control stadard except for the eccentricity of the steel pipe soil cement pile. We also discover cases where geometric tolerance is better than dimensional tolerance.

View full abstract

 We present a method for extracting and measuring cracks from digital-images of concrete surface using a simple analysis algorithm. The method is based on the shape prediction of partial domains in digital-image to determine whether the domains are regarded as cracks or not. The level-set function is used to predict the shape of partial domains in digital-image, which is also applied to measure the crack-width. We first explain the analysis algorithm for extracting and measuring cracks from digital-images of concrete surface by the shape prediction using the level-set function. Then, we apply the proposed method to several digital-images of concrete surface containing unnecessary domains such as shadows or various grimes in addition to cracks. The results offer insight into the applicability of the proposed method.

View full abstract

 We have devised a deformable retroreflector which can change the retroreflective characteristics according to acceleration by supporting one of the orthogonal three mirrors constituting the retroreflector with an elastic body and visualize the vibration from a distance. By attaching this to the measurement site of the structure and shooting a video with a camera with illumination, we developed a new remote sensing technology capable of measuring micro vibrations at multiple points simultaneously. Sinusoidal microvibration of 3 to 40 Hz and amplitude of 10 gal was applied to the prototype deformable retroreflector on a shaker and demonstrated that the vibration waveform can be reproduced from the video taken with a commercially available digital camera from a distance of 70 m . It is a robust monitoring method that can measure even if the resolution is deteriorated due to shooting from a vehicle or wind and rain due to converting the vibration into the reflected light luminance. Since the sensor does not require power supply and electric parts, it can be expected low maintenance cost.

View full abstract

 Ministry of Land, Infrastructure, Transport and Tourism (MLIT) announced Construction Information Modeling/Management (CIM) initiative and has set up a series of pilot projects to verify its value and applicability to the industry since 2012. CIM has been well recognized since its introduction. The goal of CIM introduction is to achieve further productivity gain and improvement of quality and efficiency by integrating different types of information through entire project lifecycle from design to operation and maintenance, leveraging intensive use of ICT. Current CIM research and application to pilot projects tends to just create 3D model based on a set of 2D drawings. Not much research has been seem for leveraging product models using attribute data. This paper aims to develop a prototype of 3D models of the steel bridge and substructure that equipped defined data structure of MLIT's quantity, output the MLIT's quantity calculation result from 3D Models and verified its effectiveness.

View full abstract

 This paper describes the applicability assessment of height measurement of river bank crown by Mobile Mapping System (MMS). The aim of this study is to establish MMS measurement method and extract height of river bank crown methodology from MMS point cloud. Our proposed methodology is the approach based on flatness and continuousness form of river bank crown. In order to examine our methodology, we had been carried out an actual MMS observation and MMS point cloud generation. Through our practical experiments and an actual evaluation, we confirmed that our proposed methodology can evaluate less than 10cm accuracy and continuousness height of river bank crown. It was also concluded that the MMS is the most appropriate method in terms of the efficient and effective river bank height management.

View full abstract

 It is expected to improve the whole construction process by using the CIM (Construction Information Modeling/Management). Therefore, MLIT(Ministry of Land, Infrastructure, Transport and Tourism) has studied the introduction of the CIM, and has been attempted the CIM in several projects. It is one of the issue to clarify how to add and use attribute information in three-dimensional model.
 In this paper, we reports the results of the trial for applying to add and use attribute information in three-dimensional model in the embankment project that is in the construction stage. Based on the results of the trial, we describe how to add and use attribute information in three-dimensional model for the embankment project.

View full abstract

 Ministry of Land, Infrastructure, Transport and Tourism (MLIT) in Japan is promoting the policy called “i-Construction”, which aims to improve Efficiency of construction companies. In 2016, MLIT had set up the standard for All-over Dimensional Control with Point Cloud in Public Earth Work, and had started appling this program to Earth Work.
 In this research, we surveyed the quality control levels about some Earth Work sites to propose the standard for All-over Dimensional Control with Point Cloud in Public Earth Work. In addition, we verified the accuracy of the measurement by photogrametry which process we proposed to be able to measure more efficiently.

View full abstract

 Ministry of Land, Infrastructure, Transport and Tourism (MLIT) in Japan is promoting the policy called “i-Construction”, which aims to improve Efficiency of construction companies. One of the programs which were launched first is to recommend using ICT throughout all the processes of construction, such as Survey before planning or design, measurement and inspection after Earth Work. In 2016, MLIT had started appling this program to Earth Work, then extended to Pavement Work in 2017.
 In this research, we surveyed the quality control levels about some Pavement Work sites to propose the standard for Dimensional Control with Lider in Pavement Work, then set up the process for Dimensional Control with Lider. In addition, we evaluated how much the Efficiency of Dimensional Control process would be improved.

View full abstract

 In recent years, falling rock accidents have occurred frequently, and the momentum to tackle the rock fall prevention measures is increasing. However, in lots of the falling rock prevention measures, low resolution drawings are used. Therefore, the investigating falling rocks in the field and the accuracy of the position of the falling rock generation source are poor. There is a problem of missing a place where there is a danger of falling rocks.
 In this paper, we utilize high density laser point cloud data using aircraft and create a map that emphasizes micro topography. This map is a transparent composition of the inclination amount map, the wavelet analysis diagram and the contour map together. We propose a method for extracting the falling rock origin with a high accuracy by using it. As a result of verifying the laser point cloud data in summer and winter, we could extract the falling rock generation source “cliff” by using measurement data of the longitudinal direction of the road in winter. This clarified that the proposed method is useful.

View full abstract

 In mountainous areas, it is available to construct roads of 1.5 lanes which are flexible combinations of local improvements in recent years. In construction works, 1.5 lanes are effective for temporary traffic increase of large-size vehicles. In order to appropriately and inexpensively arrange passing places, it takes trial and error by traffic simulations. Therefore, we previously developed fast evaluation method that can take probabilistic arrival interval of vehicles, as an alternative to traffic simulations. In this paper, we considered placement plan for passing places as a minimization problem of widening costs, and applied GA(Genetic Algorithm) with proposed evaluation method. As a result, in comparison with the conventional method by manual operation, it was possible to obtain an effective solution that drastically shortened the operation time from 8 hours to 4 minutes, and also reduced both the widening costs and the average waiting time. In addition, using multi-objective optimization, trade-off relationship between the widening costs and the average waiting time was clarified.

View full abstract

 A long period of time has passed since the design of the steel bridge has shifted to Guideline based design. But, creating materials to obtain consultation / confirmation / approval with the client in detailed design depends on the experience of the engineers. There are many solutions where the sum of the material cost and the production cost becomes minimum according to the change position of the cross section. Therefore, in order to smoothly carry out consultation and approval, it is necessary to select and explain and compare alternatives appropriately.
 In this study, we proposed a tool to retrieve alternative proposals by GA which minimizes the sum of material cost and production cost for each number of parts. It was targeted the steel continuous girder bridge of the MLIT (Ministry of Land, Infrastructure, Transport and Tourism) type and former JH (Japan Highway) type. It showed that the method of extracting, evaluating and comparing multiple alternatives that minimize the sum of material cost and manufacturing cost in determining cross-sectional change position is effective for detailed consultation and approval.

View full abstract

 The Ministry of Land, Infrastructure and Transport is working to introduce and disseminate CIM (Construction Information Modeling / Management) in order to make the construction production process more efficient. CIM is aiming to achieve cooperation between the construction production process and 3D models. This cooperation will facilitate information sharing among participants, improving productivity. The effectiveness of CIM is confirmed at the Design and Construction from the results of field trials. However, maintenance stage on-site trials are still not done. Therefore, through this research, the effectiveness of CIM at the maintenance stage was confirmed by 4 procedures. First, we developed a prototype of a CIM information managementplatform system intended for use at the maintenance stage. Secondly, we examined a method of utilizing the prototype system at the maintenance site. Thirdly, we conducted a trial of the prototype system at the maintenance site. Finally, we evaluated the effectiveness of a CIM information management platform for maintenance.

View full abstract

 In the "Periodic inspection rocedures" of various road structures formulated in June, 2014, it is mandatory to make the regular inspection for existing road structures once every five years. For this reason, technological development for an efficient maintenance management was pointed out as an urgent task. “Mobile Imaging System”, which was developed few year ago, measures road structures by using digital video camera and laser scanner mounted on a specialized vehicle. In addition, the measuring precision is high. However, in many cases the operation cost is expensive. Therefore, it is difficult for local government with budget limitation to apply it to the regular inspection. Authors has developed a reasonable “Mobile Imaging System”, that is an mountd on a maintenace patrol car, measurement accuracy same as a specialized vehicle, decomposable and has developed an image stiching program for panoramic images.

View full abstract

 Large-scale earthquakes are often accompanied by simultaneous multiple fires. In addition to the collapse of buildings due to the earthquake, inundation by the tsunami, etc., it is necessary for local governments and residents to understand the risks of fires properly. It helps not only the initial fire fighting and evacuation by residents in case of fires, but also the town planning against disasters and actions by fire stations.
 The spreading fire analysis can be used for the evaluation of the earthquake fire risk, but careful attention should be paid to the generalization of the analysis results which changes greatly depending on the conditions such as allocated fire origins.
 In this study, a new earthquake fire risk assessment index is proposed considering the spreading time of fires, by conducting the spreading fire analysis with all the buildings as fire origins in the target area.

View full abstract

 Ministry of Land, Infrastructure, Transport and Tourism announced i-Construction aimed at increasing productivity in the infrastructure maintenance in 2015 by introducing 3-dimensional data using satellite positioning. The accuracy of the satellite positioning system is high in the horizontal direction, however, low in the vertical direction. The purpose of this study is to compare the vertical accuracy by GPS and QZSS to the vertical accuracy by GPS alone. As the results, it was shown that QZSS contributed to keep high vertical accuracy for long time when the elevation angle was more than 80 degrees.

View full abstract

 The atmospheric CO₂ concentration has been on the increasing trend since Industrial Revolution. The annual average CO₂ concentration in 2015 exceeded 400 ppm and it updated to the highest value. Ibaraki University has measured CO₂ concentration in the living area of Ibaraki since 2007. So far, we confirmed the fluctuation pattern of CO₂ concentration is different for each region and is affected by the amount of vegetation around measuring points. Then, our research group proposed the new environment evaluation index focusing on this feature. However, in the past research, we did not fully verify the effectiveness of the index. In this paper, we aimed to calculate the index from CO₂ concentration for about three years and to investigate correlation with geographical information (e.g. land use and population). As a result, we found the high positive correlation with vegetation in summer and positive correlation with human activity in winter. Therefore, we conclude the nature of the index changes according to the season.

View full abstract

 The ability to obtain a pattern of cracks and their dynamics, along with ability to acquire and analyze data of cracks increase in length and width, provides an effective way to inspect and monitor the condition of concrete structures. Using conventional methods, until nowadays crack inspections have been performed manually, with sketches, which reflect positions and shapes of cracks, and which have been drawn by hand. As a result, the changes cannot be identified and reflected accurately, as far as the data obtained with this method is typically imprecise and to a considerable degree dependent on the individual skills of the inspector.
 In this research, we propose a new method for inspection of concrete cracks by combining digital image processing and KUMONOS, which is a remote concrete crack inspection system using a total station. In this method, as-built dimension of the structure and several cracks on the structure are measured by KUMONOS. Image correction based on the as-built dimension obtained by KUMONOS leads to accuracy increase. Furthermore, using the data of crack widths as a reference value, the width of the other cracks on the digital images can be accurately measured. Consequently, the new method increases the accuracy in image correction and measurement maintaining the efficiency of crack inspections.

View full abstract

 ALOS-2/PALSAR-2 data is not affected by the weather, it is expected to know the damage of the dis-aster because it observes the sky over Japan twice a day and night. Extraction of the flooded area using the backscatter coefficient image obtained by SAR data has already been studied for practical use. However, in urban areas and forest areas, there are few places where the ground surface can be directly observed, and it is difficult to know the flooded area with only SAR data. On the other hand, there is a flooding analysis that estimates the flooded area by hydraulic method.
 In this study, a method of extracting a flooded area combining SAR data and flooding analysis was proposed. And it applied to actual flood damage. As a result, it was shown that the proposed method can extract flooded area in urban area.

View full abstract

 The road tunnel periodic inspection manual indicates that the frequency of the periodic inspection is once every five years. To reduce working times and costs for tunnel lining inspection, an automatic inspection system is required. To achieve this, the authors have developed an inspection system for concrete tunnel lining based on image stitching and processing techniques. In this paper, the image acquisition device developed by authors was used in order to take continuously images of the tunnel lining surface at 30km/h speed. The device consists of six video cameras and three illuminators arranged optimally on a steel framework to ensure the quality of the captured images. Moreover, the device is mounted on an inspection vehicle to record the entire tunnel lining surface image for practical inspection application. In previous studies, the authors proposed the automatic image stitching method based on a local search algorithm in order to generate panoramic images of tunnel wall from the scanned images. However, the stitching method needed information of the vehicle speed and the imaging device such as frame rate. Therefore, in the paper , the gradient of intensity difference distribution between two images for image stitching was applied to find appropriate matching points without information such as vehicle speed.

View full abstract

 In July 2014, the inspection of bridges was obligated. It is conducted one every five years. Inspection method of cable-stayed bridge cables is largely divided into two ways. The first way is the visual inspection using aerial work platforms. In the case of the platforms, the height from the bridge surface is limited to about 30m. In addition, traffic regulations is required for the inspection. The second way is the inspection using climbing techniques. The climbing is difficult to ensure the safety of inspectors.
 To overcome these problems, the authors have developed an automatic inspection robot for the cables of cable-stayed bridges. The purpose in the study is to develop an inspection record system for the robot. Concretely, the system takes continuously scanned images of cable surface by using USB cameras, miniPC, and lighting devices. Then, it generates a panoramic image of cable surface. However, the existing system could not create an accurate panoramic image if the existing system robot rotates while recording. Therefore, in the study, the author developed the image combining method for taking account of rotation of the robot by using information of angle measured by clinometer mounted on the robot. Furthermore, the system was applied to the inspection of 7 meters cable tube in order to demonstrate the accuracy of panoramic image.

View full abstract

 In this study, a screening method to detect any damage to a lighting pole using an acceleration sensor is considered. The spectral shape near the primary vibration component of a lighting pole is focused. The selection of an acceleration sensor and the invention of an vibration method is considered, where this spectral shape can be obtained stably. By the relative comparison of spectral shapes by many lighting poles, the shapes of spectral disturbance are understood, and the relationship between the shapes of spectral disturbance and damage types is considered. Finally, the accuracy of damage detection of this screening method is clarified.

View full abstract

 The Ministry of Land, Infrastructure and Transport promotes i-Construction to improve productivity at the construction sites. In this approach, utilization of 3D data by Laser Scanning and SfM (Structure from Motion) with unmanned areal vehicles (UAVs) is one of the main axes. However, the laser scanner easily causes data loss due to constraints on limited installation locations, and SfM cannot obtain the actual dimensions. In this research, we propose a complementary integration of the scanned data and SfM data, considering the distribution of the 3D point cloud. Performing SfM using the image taken by the calibrated camera equipped with the scanner and the aerial image set is capable of reconstructing the scanner viewpoints and the scene geometry in an identical 3D space. The reconstructed 3D scene by SfM can be overlaid on the scanner's camera image, whose pixels are associated 3D point cloud coordinates. Then the SfM data is associated with the point cloud by one-to-one correspondences. The point cloud is then divided into lattice-like areas to resample with a limited number of points. The points to be used for finding the trans-formation to align the SfM to the point cloud, are selected evenly dispersed in the 3D space, by considering the distribution of the non-empty lattice-like areas. In this way, the coordinate transformation can be found by explicit calculation without iterative computation. Our implementation achieves automatic registration capability and the experimental results showed that the registration can be stably done with 2.0 cm order error.

View full abstract

 In the construction work of the mountain tunnel, it is a problem that workers are injured due to disaster caused by fall of rock. In ordet to prevent labour accidents , it is important to monitor deformation of tunnel face during work constantly. However, it has been difficult to detect the deformation by visual monitoring of the supervisor. Therefore, we have been developing tunnel face deformation monitoring system using background subtraction which is a method of image processing to detect moving objects. In this study, in order to make it possible to monitor the deformation by image analysis, we examined the applicability to the site from analysis results of videos moving rock and that excavating tunnel face. As a result, it was found that it is expected to detect the deformation that is difficult by visual monitoring, and to detect only the deformation by removing moving objects other than deformation. Finally, we indicated possibility that the risk of accidents is reduced.

View full abstract

 This paper presents a visualization system based on the marker-less augmented reality for unsteady water environmental and disaster simulations. The marker-less method using a landscape image is employed for the positioning method for the virtual information and real images. The present system is applied to typical water environmental and disaster problems in river and coastal region in order to investigate the applicability of the method. The present method is to be a useful tool for the visualization in case that a lot of feature points are located at the simulation area.

View full abstract

 In order to promote congestion countermeasures, it is important to grasp the origin of the traffic jam and the extent of the influence of traffic congestion. In the previous research using VICS and car probes, a method using a section unit of a digital road map as a minimum unit has been used. In this research, a method of analyzing in divided section unit such as not being influenced by DRM section length by processing the travel trajectories (traveling position and time) in the ETC 2.0 probe data was proposed. Comparing the evaluation results of the conventional method and the proposed method, in order to evaluate bottlenecks and influence ranges more precisely and properly, it was shown that it is effective to utilize the ETC 2.0 probe data.

View full abstract

 Many research and development has been done for flow rate measuring methods of rivers and so on, and they have been put to practical use both in Japan and abroad, but as far as we know there is no way to automatically and continuously measure river water from low to high level water. We have developed a sensor called a point current sensor, aiming at a system which can combine small radios, networks and proprietary web servers to automatically measure river flow and can deliver in real-time. Since the point current sensor is compact, inexpensive, consumes less power based on the simple principle using a reed switch. It is suitable for combination with a wireless device and is unnecessary to construct a large-scale cable under the river bed, therefore it is easy to install. In this paper, we explain the idea of the flow rate measurement system, the principle of a point current sensor and the solution method of the problem found in the field experiment so far and also show the development situation of the whole system.

View full abstract

 In road traffic analysis, it has become possible to make a qualitative analysis by extracting road users' comments from microblogs, in addition to a quantitative analysis using the data obtained from diverse instruments installed on roads. Bearing utilization in abnormal situations in mind, it is necessary for the road administrator to be able to extract only highly authentic posts out of various posts instantly; however, since anyone can post microblogs freely, the authenticity of each post is unclear. This characteristic is harmful in utilizing microblogs in abnormal situations. In this research, a method for extracting the posters of authentic articles related to traffic phenomena was devised in terms of the poster's attribute and posting tendency, and its usefulness was verified.

View full abstract

 The point cloud data generated from UAV photogrammetry are put to various uses such as management of work amount completed and as-built conditions in earthwork, as well as prompt grasp of the situation during disasters. Their accuracy affects measuring conditions of UAV and analytical parameters of the photogrammetry software; therefore, it is necessary to specify it appropriately depending on the situation of the sites. Existing researches have proposed methods for ensuring the required accuracy by analyzing measuring conditions. However, even when the measuring conditions are appropriate, it has a problem that the accuracy as well as costs in time and man power may vary depending on the difference in analytical parameters. Another problem is that when it is impossible to set up ground control points in the circumstances during disasters, suitable analytical parameters cannot be set up. This research proves the relationship between time and accuracy depending on the difference in analytical parameters, proposes an analytical method for UAV photogrammetry that is assumed to be utilized during disasters and confirm to apply to the real disaster site.

View full abstract

 Many natural disasters such as earthquakes and floods occur in Japan. In the case of earthquakes outside working hours, at night or on holidays for example, officials of disaster management bureaus of government agencies need time to return to their offices. It is, therefore, necessary to construct an automatic system which rapidly shares disaster information among such officials.
 The Ministry of Land, Infrastructure, Transport and Tourism (MLIT) has developed and been utilizing CCTV (Closed Circuit Television) cameras to survey infrastructure on roads and rivers and to support disaster management. Officials of disaster management bureaus of the MLIT attempt to check as many CCTV cameras as possible to clarify damage to infrastructure. Under such circumstances, they have no time to rotate the CCTV cameras to obtain a broad panoramic view of their surroundings. To support early-stage governmental decision-making after an earthquake, the National Institute for Land and Infrastructure Management (NILIM) has been studying the construction of a system that will automatically select the CCTV cameras in areas where the measured seismic intensity exceeded a preset level based on municipal unit seismic intensities and rotate the corresponding CCTV cameras to render panoramic images of their surroundings.
 To introduce the system, we have to clarify the time officials of MLIT disaster management bureaus need to check panoramic images. CCTV cameras satisfy a common specification set by the MLIT to maintain their quality throughout Japan. Therefore, we must also clarify the method of rotating each kind of CCTV camera to completely pan its surroundings.
 We confirmed the behavior of the system and performed trials to measure the time it takes for 10 CCTV cameras along Tsurumi-River to completely pan 180 degrees. Under this condition, we clarified that rendering panoramic images is failing when we spend no more than 8 seconds.

View full abstract

 Recently, large-scale typhoon disaster beyond a city occurs frequently, for example Kanto-Tohoku heavy rainfall in September 2015. As one of a soft disaster prevention measure for correspond to such disaster, making of a timeline plan is promoted. However, preliminary learning and practice is needed for emaergency action along a timeline plan. The aim of this study is to develop an action support database for evacuation study of a emargency action. This database is stored emergency action information in the past typhoon disaster. Effective emergency action is considered using this database and it's analyzed about a problem of emergency action in the past. In addition, the learning method is proposed using action support database for evacuation study in typhoon disaster.

View full abstract

 In the past years there has been an extensive effort to develop an automated crack detection method by image processing to improve inspection and evaluation process of concrete structures. However, these methods are not yet accurate enough due to the difficulty and complexity of the problem. Especially, the mold mark is misjudged as the crack because image characteristics are quite similar to each other. To solve this problem, this paper proposes the method which distinguishes cracks and mold marks properly by convolutional neural network which is a type of deep learning. Then, accurate classification is archieved by the random forest method with considering image characteristic related to pixel value and geometric shape. The accuracy of developed method is investigated by the photos of concrete structures with lots of adverse conditions including not only the mold mark but also shadow and dirt, and it is found that the method can extract the crack region with high accuracy.

View full abstract

 Recently , rise in temperature of the urban area by the heat island effect becomes the problem under the influence of global warming and urbanization. Therefore, Tokyo Metropolitan Government was aimed for temperature reduction and established the regulations about protection and the recovery of nature in Tokyo in 2001. These regulations are aimed for improvement of the green coverage ratio and demand the presentation of the afforesting plan for a company developing the applicable structure construction. Temperature reduction caused by the roof tree planting attracts attention, and increase and transpiration of the quantity of absorption of carbon dioxide are expected. The purpose of this study was to examine the extraction method of the roof top greening area of Chiyoda Ward by the normalized vegetation index value by using the multispectral satellite observation information of WorldView-3 satellite data. The extraction accuracy of the roof top greening area of the sunshines region and the shades region was appropriately evaluated by combining using the GIS data of the building shape was shown. It was considered that this technique was effectively used from these results to evaluate the roof top greening in the future.

View full abstract

 In order to improve the efficiency of social capital management, it is required to develop an information infrastructure in which information such as facility information and inspection history is stored in a database, and to utilize the accumulated information. Although it is possible to construct a deterioration prediction model from such information, the highly accurate prediction for individual management sections is difficult because unobservable variables mutually influence. On the other hand, if it is possible to estimate the range within which the prediction can be accommodated even with uncertain predictions, it can be expected that it will be useful for the manager's decision making.
 In this research, we proposed a method to estimate confidence intervals and prediction intervals by considerably reducing the amount of computation by efficiently applying the bootstrap method to the deterioration prediction model using neural networks. The proposed method was applied to the rutting depth of the road surface condition survey history which is the inspection data of the pavement surface, and the estimation accuracy of the prediction interval was verified.

View full abstract

 A large number of highway bridges built during the period of high economic growth are now aging simultaneously, requiring appropriate maintenance. However, the design drawings and completion drawings have been disposed of for some of these highway bridges, or they don't agree with the present conditions. As this makes maintenance difficult, recovery of as-built drawings is required. For this problem, the authors have attempted automatic generation of road alignment drawings by measuring overpasses using MMS. However, the proposed method only verifies the accuracy of the drawing generated using one type of MMS's point cloud data. But, We did not apply the method to others of MMS's point cloud data which have greatly different properties such as the type of laser scanner or the direction of installment. This research verifies the versatility of the proposed method by comparing the result of extracting feature points of the overpass and the accuracy of road alignment generation using two types of MMS's point cloud data having greatly different properties from each other.

View full abstract

 Recently, a lot of natural disasters have occurred frequently in Asian region. Especially, the earthquake off Sumatran suffered a tremendous damage by tsunami. Ten years pass from struck respectively, and the revival had been advanced in Sumatran disaster. It is important to analyze the revival for the development of the disaster area in the future. The authors had already the satellite remote sensing in the time series by a different resolution effectively analyzed the revival situation for a long-term. However, based on the original land covering environment of the disaster area there is not enough analysis or consideration after the disaster revival about and redevelopment situation. In this study, investigated the change of land cover before the disaster using a plurality of high resolution satellite images in time series, analyzed long term changes of land cover classification about revival process. Moreover, the changings the land cover of each pixel were shown by calculating the difference of the land cover classification image at two time. As a result, it was evaluated that the restoration characteristic of the city from struck was more effective.

View full abstract

 The purpose of this research is to develop a low-cost traffic flow measurement system. To achieve the objectives, we have utilized vehicle detection technology based on HOG and SVM. We have experimented traffic flow measurements on the multi-lane public road using the developed system. The results have been compared with the background subtraction method and YOLO (You Only Look Once) that is based on convolution neural network. Our system has demonstrated both real-time performance and better false detection and undetected rates.

View full abstract

 The network models of path-finding systems for pedestrians are generally defined based on linear models similarly to the automotive navigation systems. However, pedestrians move around avoiding obstacles on a planar area, and a network model that takes these into consideration is necessary.
 In this paper, we conducted literature survey on the graph methods in the motion planning and selected a visible graph and a grid graph as suitable models for pedestrian movement. We defined a hybrid model that takes advantage of both graphs and implemented it in the path-finding application. Also, we evaluated the multiple cases with different movement costs in the region and clarified the usefulness of the hybrid network model.

View full abstract

 Recently, due to the frequent heavy rainfalls in short-time, inundation damages in underground spaces becomes higher risks in urban areas. Although advanced prediction techniques are available to analyze the inundation damage relative to heavy rainfall patterns, the method to convey the predicted inundation information to the onsite people is not yet established. Since the positioning services based on GPS do not function in the underground, location-oriented services are not applicable for notifying the inundation hazardous information to the people, most likely smartphone users, in the underground space. In this study, we propose a markerless AR method using 3D point cloud data captured by laser scanner instead of GPS. By using the natural feature points of the multi-view images rendered from the colored point cloud data, it is possible to estimate the current position of the user, via a photo image taken by the user's mobile terminal on the spot. The colored point cloud can be 3D coordinates references and rendered images and user's photos are associated with the pixel positions on the natural features of robust key-point features. Then the camera position and orientation can be estimated relative to the world coordinate in the real dimension of the point cloud and predicted flooding condition is superimposed on the user's photo in AR representation. Moreover, it is possible for users to intuitively understand the flood depth by using flooding expression reflecting the relationship of water surface and objects based on the 3D models of the target space generated from the point cloud, this paper shows the effectivity of our method by experimental implementation at an actual railway station facility in underground and inundation simulation results based on assumed heavy rainfalls of past data in Osaka city.

View full abstract

 This paper presents a railway noise evaluation system using virtual reality technology. The computation of noise level is performed by the geometric acoustic theory. In order to realize the stereo sound field, we introduced the ambisonics based on thespherical surface function expansion to realize the stereoscopic sound field. In order to consider car composition, we implemented the sound source data which is classified into motor car and non-motor car. The computed results are compared with the measurement data in both real space and virtual reality space. The present system is shown to be a useful tool for planning and designing tool for railway environments.

View full abstract

 This article describes the method to produce the 3D model of a site by the photographic surveying using UAV and TS, which improves the performance in the construction area. By this method, the number of the ground control points can be reduced. It has been developed in combination with the three basic technologies; distance and angle measurement using TS, photographic surveying and synchronization technologies of each sensor.
 The demonstration experiment has shown the precision of 3D surface model to be less than 50 mm. With the demonstrated precision, the method will be applicable not only to pre-construction surveys or as-built surveys in i-Construction, but also to measuring the damage level of a site in large-scale landslide disasters.

View full abstract

 Digital road maps and large-scale road maps (fundamental geospatial data of road) are drawed up and updated using completion drawings for road construction. To promote a series of measures including i-Construction makes it possible to expect circulation of three-dimensional (3D) data that represent shapes and characteristics of roads in detail in addition to the conventional two-dimensional data such as completion drawings for road construction. Consequently, technological development for reusing 3D data is becoming more important. 3D data are likely to maintain the shapes, elements, and attributes of road features. This makes it is expected that 3D data will be applied to draw up and update of fundamental geospatial data of road as well as road maps such as digital road maps. On the other hand, their implementability have yet to be elucidated fully.
 In this study, devised methods for drawing up and updating of road map using 3D data that expected to circulate in the movement such as computer-aided construction, CIM and i-Construction. Additionally, verified the usefulness of the above methods by conducting case studies.

View full abstract

 In the near future, it will become possible to obtain probe data with high positioning-accuracy, which enable to identify traffic lanes. For this reason, it will be raised expectations to the use and the realiza-tion of advanced road traffic analysis. In order to realize the lane-by-lane road traffic analysis, the large-scale road map composed of traffic-lane network data in addition to probe data. However, existing large-scale road maps are focused on the use for driving support services such as route guidance, and it is not made clear whether they are suitable for road traffic analysis.
 In this study, authors discussed the requirements and measures for large-scale road maps by specifying use cases of the lane-by-lane road traffic analysis based on research results on use cases of existing road traffic analysis. Additionally, authors extended the existing large-scale road map in line with above measures, and conducted case studies by the high-accuracy probe data and the extended large-scale road map.

View full abstract

 In the road surface inspection, the presence or absence of cracks is judged by visual interpretation of the road surface image taken. This process occupies about 20% of the whole process, which is a bottleneck in production. In this research, we developed a classifier that classifies cracked images using machine learning techniques, and improved efficiency by replacing visual interpretation with machine interpretation. The road surface image of the road managed by municipalities was divided into 50 cm squares, learning was carried out with about 140 thousand images, and the evaluation was carried out with about 15 thousand images, and the correct answer rate was 90.6% . Next, using this same classifier, judgment on the presence or absence of cracks was made on about 6 million images of prefectural control roads, and crack rate was calculated by aggregating every 20m and 100 m. The correlation coefficient between the value obtained by visual interpretation and the value obtained by machine interpretation was as high as 0.875. The time spent on machine interpretation was reduced to about 40% of the time required for visual interpretation. In addition, the actual working time of human beings has been reduced to about 2%.

View full abstract

 ICT construction machinery used under the machine control or machine guidance systems at construction sites obtain history data on their behavior and traveling. i-Construction, which is promoted by Ministry of Land, Infrastructure, Transport and Tourism, applies these history data to calculation of progress payment. It can be expected that circulation of history data will develop in the future.
 This study proves applicability of the history data to as-built management as a supporting measure to help further application of history data. This paper devises a method for generating 3D models using history data and a method of as-built management using 3D models. Then it verifies the usefulness of the devised methods by conducting a case study.

View full abstract

 Due to the spread of three dimensional measurement devices such as MMS, ground based laser scanner and camera mounted UAV, the technology concerning practical use of the measured point cloud data has attracted attention. In addition, as a part of the measures of “i-Construction” promoted by the Ministry of Land, Infrastructure, Transport and Tourism, the as-built management, and the supervision and inspection guidelines using point cloud data are published, and it is expected to increase the utilization scene of point cloud data in future. In order to effectively utilize point cloud data, it is necessary to extract road objects from them. Therefore, many previous researches analyze the shape and position information to extract objects. But it is difficult to extract road objects comprehensively. Thus in this research, we propose a method to extract road objects from point cloud data referring to the road completion drawing on the road construction completion reports.

View full abstract

 In Japan, the airborne laser has been carried out for the class-A rivers nationwide for the purposes of verifying the cause of the flood associated with climate change and torrential rain, and suppressing the damage. In the previous research, by recognizing crosswise and continuously the topography obtained from point cloud data, it is possible to generate a three-dimensional model on river with high accuracy. Meanwhile, as the Ministry of Land, Infrastructure, Transport and Tourism promotes full use of ICT, it has developed specifications and standards related to three-dimensional model called CIM. As one of these standards, there is a three-dimensional design data exchange standard that establishes the data structures and formats of electronic delivery data relating to river projects. Thus in this research, we propose a method to detect objects on river embankment from point cloud data. This makes it possible to generate three-dimensional models that can be compared with three-dimensional design data exchange standard expected to become widespread in the future, and promote the full use of the accumulated point cloud data.

View full abstract

 Three dimensional building models were made by SfM and TLS. The mean values and standard deviations about measured eight long sides, eight short sides and twelve heights were calculated in order to verify the accuracy of the three dimensional building models by SfM and TLS. The surveying by TS was also conducted at the same lines. The relative errors, assuming that the mean values by the TS were true, and the standard deviations of the three dimensional building models were calculated in order to verify the accuracy of SfM and TLS. The scale of three dimensional building model by SfM were corrected, by focusing on the TS and assigning the actual measurement value by the TS to the height of the three dimensional building model by SfM, in order to improve the accuracy of the building model. As the result, the order of high accuracy was “TLS”, “Corrected SfM” and “SfM.” It was also found that the SfM could decrease the relative error from around 5% to around 1% and improve the accuracy to the extent of around 1.5% difference from the relative error of the TLS, by assigning the actual measurement value by the TS to the model and correcting the scale.

View full abstract

 Currently, in the introduction plan of LRT being advanced in the Haga town and Utsunomiya city of Tochigi prefecture, the whole line is going to be newly constructed routes. Considering the current situation of cities that already introduced tramway where contact accidents with cars are occurring, there is a high possibility that a contact accidents with cars will occur after the introduction of LRT. Therefore, we focused on the scenario where contact accidents are likely, namely, turning right at an intersection, and developed VR contents that can replay the situation using Multi-platform the after the introduction of LRT. In this VR contents, we have faithfully reproduced the road traffic phenomenon that is close to the reality based on the plan of LRT using an integrated development environment of traffic flow simulator and game engine. Also, in order to confirm the completeness of VR contents, we have conducted a monitoring test targeting citizens using inexpensive and easy to carry VRHMD. As a result, some amendment/improvement points became clear, such as the need for weight reduction of VR data as well as addition of additional functions etc. and we have examined a few implementation issues.

View full abstract