BUNEN KYO

The authors have developed "xG-1", a sports sensing unit capable of visualizing the performance and flow of play while recognizing each player in field sports such as soccer, rugby, and American football. Utilization of this sensing unit in the field of sports makes it possible to provide scientific data-driven services with the aim of transforming subjective coaching methods based on experience into new coaching methods based on data. This system consists of sensor devices and software, utilizing real-time kinematic (RTK) positioning, which is a surveying technology. Therefore, the system not only provides highly accurate positional data in real time but also visualizes the overall performance of the entire team, including formation analysis. This allows one-stop management of the overall team and individual performance, as well as efficient information management of the player's exercise load. "xG-1" can support sports businesses by providing highly extensible API and SDK and linking data among end users and systems. In the future, we aim to develop a global open data platform.

Research on tracking and performance analysis of athletes using video images has been actively conducted with the aim of improving athletes' competitive performance. However, when filming plays in field sports, it is difficult to capture the entire field with a single camera without filming from a specific point, such as a spectator's seat on the corner side, because the field is long sideways. Even if the entire field is captured, the players at the back of the field appear small, making analysis difficult. Other issues include the fact that since many coaches and analysts film where the play is progressing, it is hard for them to track the ball seamlessly when the position of play changes significantly depending on the position of ball. To solve this problem, we develop in this study a technology to automatically generate panoramic video images that cover the entire field by using two video cameras. Using this technology, we aim to generate panoramic images of the entire field that makes it possible to surely measure and analyze all players and the ball.

In recent years, effective safety management measures based on the Internet of Things (IoT) have been established to prevent industrial accidents in construction sites. In response to these efforts, technologies that use object detection methods to obtain the location information of workers and construction vehicles can contribute toward safety management. However, it is difficult to detect workers and construction vechiles from video images by object detection methods in construction site where dangerous areas are constantly changing due to their coexistence. Therefore, the application of deep learning can be considered, but in order to detect them accurately, it is necessary to update the detection model specialized for construction sites. However, it takes a lot of cost to manually create a training data for constructing model. Therefore, in this research, we propose an automatic generation method of training data for deep learning to realize high-precision for detecting workers and construction vehicles in construction sites. Experiment was performed with application of the proposed system to the video data at construction site, and its usefulness was confirmed.

In Japan, owing to the steadily declining working population, innovative technologies that improve productivity in construction sites using i-Construction are urgently required. This necessitates the tracking of not only construction vehicles but also workers using technologies such as AI and IoT. Against this background, many technologies that were able to track both workers and construction equipment on construction sites had been developed. However, challenging subjects for automating tracking tasks remained as those technologies were applied in sections where extended and heavy occlusion occurs frequently. Therefore, an identification method that matches appearance features of workers and construction vehicles before and after occlusion is proposed in this research, thereby contributing to the development of tracking technologies that enable successive tracking through occlusion sections. Experiments conducted using videos from actual construction sites showed that the method was capable of restoring tracking progress in occlusion sections by identifying each worker and construction vehicles individually before and after each section.

The authors have aimed at detecting and tracking players using video cameras in field sports. However, there is a problem that the detection accuracy is lowered at the place where players have an occlusion. Deep learning is a solution to this problem, but it requires the manual creation of large amounts of training data which is costly. In this research, the authors propose a method for automatically generating training data to build a detector that can identify players with high accuracy for field sports, and a method for updating this detector to input it to various videos.

Studies on tracking athletes by image processing are being made for the purpose of utilizing data in sports using ICT. However, it is impossible to correctly obtain the position information of athletes in the types of games such as American Football in which occlusions occur frequently due to the athletes’ close formation on the field. Therefore, it is a problem that the moving track of an athletes is divided, which makes it difficult to make accurate tracking. In this research, we propose a tracking method for an athlete who is robust against occlusion by using the moving tracks obtained from respective video images shot from multiple points to make them complement each other. Using this, we aim at tracking in regular games to obtain information including that about the opponents that cannot be obtained from the device with GNSS sensor.

Japan Sports Agency aims at supporting distinguished performance of national members of Japan from a scientific aspect in the prioritized policy concerning improvement in international athletic ability. Focusing on the field sports, we have been developing a system for visualizing athletes’ plays using the GNSS sensor. In particular, we have been performing a research on matchup analysis of pass plays with a focus on American football. The aim of the research was to decide whether a pass is completed or incompleted by deep learning, using trajectory images of matchup of QB, WR, and DB. However, since only the track information of the finished plays was used, it failed to obtain information perceived during the play, for example, prediction of the completed pass probability during the play, or information that enables directing timely choice or modification of action from the side line during the game based on the prediction values. In this research, we attempt to predict completion probability during plays by using track information that takes into account the time term. In experiments, we made analysis of quick passes, short passes, and long passes to demonstrate its usefulness.

In our country, policies regarding sports are actively advanced towards the Tokyo 2020 Olympic Games. One of those policies, ”Sports x ICT” considers effective methods of utilizing ICT (Information and Communication Technology), such as development of measurement instruments, measurement and visualization of data, and proposals for new services in the field of sports. Against this backdrop, we have been developing the visualization system for American football games using terminal devices included GNSS and acceleration sensor. Using that system, we confirmed to grasp the effective information that are the individual condition and the motion analysis of American football players against not only players but also the college football leaders. But, in our existing research, we could not realize the strategy analysis of games that are to select the play calls depending on circumstances, to make a prediction of successful ratio, and so on. Then, in this research, we apply deep learning to the matchup analysis of pass play included offence and defense players, and then verify whether it is possible to infer the success or failure of a play from it.

我が国では，2020年の東京オリンピックに向け，日本代表選手の活躍や競技水準の向上を目的としたスポーツ情報科学に関する研究支援が求められている．そのため，昨今では，スポーツ分野に特化した計測機器が新たに商品化され，それらを用いた計測データの可視化システムの開発や，戦術分析のための応用研究が試行されている．しかし現状では，選手個人のリテラシーを指標化することにとどまり，チーム全体のパフォーマンスを多様な観点から分析できるような現場指向型システムにはほど遠い．そこで，本研究では，サッカーを対象とし，選手個人だけでなく，チームのフォーメーション，選手間の間合い，連携度合，試合の局面・流れなどを確認できる現場指向型スポーツ可視化システムを考案する．この実現に向けて，まず，既存研究の調査と専門家へのヒアリングを通して，必要な可視化項目を整理する．次に，可視化項目に基づき，利用場面を考慮したシステムの各機能を構築する．そして，提案システムを用いて，実試合における選手個人やチーム全体のパフォーマンスを分析し，指導者に新たな気づきを提供できるかの視点から実用の可能性を検証する．最後に，提案システムの新たな展開の可能性について論述する． In Japan, to prepare for the 2020 Olympic Games, additional research in sports information sciences is necessary to improve the performance and competitive level of athletes. For this reason, new specialized measuring instruments that are uniquely applicable to the field of sports are available for purchase. The development of data visualization systems and the application of tactical analytic research using these instruments, are also being attempted. However, current instruments and methods are only capable of indexing athletes' individual performance levels. They are inadequate as field or team-oriented systems that can analyze the overall performance of a team from various perspectives. Therefore, in this research, we propose a coach-oriented sports visualization system for soccer that can confirm not only the individual capabilities of the players, but also the optimized team lineup, distance between individual athletes, competitive fitness, and also analyze various phases of a game. To develop this system, first we organized the items that needed to be visualized by studying existing research and interviewing soccer specialists. Second, we considered various applications for applying this system, and then created every function based on the visualization items. Third, we analyzed the performance of individual athletes as well as the entire team in actual competition to provide new ideas and supporting data for instructors based on verifiable, practical possibilities. Finally, we summarized the areas of potential improvement for this proposed system.

In our country, policies regarding sports are actively advanced towards the Tokyo 2020 Olympic Games. One of those policies, “Sports × ICT” considers effective methods of utilizing ICT (Information and Communication Technology), such as development of measurement instruments, measurement and visualization of data, and proposals for new services in the field of sports. However, since attempts to apply ICT to sports have just started these days, it is difficult not only for the leaders in schools and local sports clubs but also for the staff who possess advanced expert knowledge about sports to make full use of ICT as desired. In this research, an investigation is made into the trend of studies related to “Sports × ICT” with a focus on field sports, and a visualization system for a positional play analysis is developed to grasp the present state of team and to grope the methods of practical use. Then, its practicality is verified through motion analysis of American football players using the measuring terminal device to be combined the statistical methods, based on whether it is capable allowing the college football leaders such as managers and coaches to perceive something new.

Our country is promoting supporting activities for training players and improving the competitiveness in preparation for international sporting events including Tokyo Olympic Games. Especially recently, one of the supporting activities at present is to utilize high-performance sensing devices for analyzing tactics by measuring the positional data of the players and digitizing their performances. However, sensing devices have not been put to practical use yet because it is hard for the players to wear them during games of various kinds of sports. Therefore, technologies for identifying players by image processing from the video images are the mainstream of the sports information fields. However, they have a problem that the identification accuracy decreases because lots of players are occluded into image data. Thus, in this research, we use multi-cameras from single viewpoint to identify American football players with high accuracy and specify their positions by photographing occlusion points with high zoom magnification, and then applying deep learning of the OpenPose. Finally, we embed the identification results of the OpenPose to the video images taken with low zoom magnification and try to identify players and to analyze their positions.

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.

Many highway bridges constructed during the high economic growth period have become deteriorated. So their maintenance is required. However, they are facing difficulties in maintenance because their drawings prepared at the time of design or completion were made in paper media, many of which have already been disposed. There is a research that attempts to automatically generate SXF (Scadec data eXchange Format) drawings of a highway bridge to use for its maintenance, using point cloud data obtained by MMS (Mobile Mapping System). In the previous research, the authors use positions of joints on a bridge to divide a point sequence into each span and generate their alignments of the highway bridge. However, when it is unable to extract a joint, or when the start and end points of the alignment are placed at a short distance from the positions of joints on a design drawing, the generated clothoid curve will lose continuity with straight line and circular arc. Thus, this research proposes an analysis method for calculating the start and end points of straight lines, clothoid curves, and circular arcs in a plane and straight lines and quadratic curves in a longitudinal section to allow each alignment to connect with each other naturally. This allows generating the SXF drawings of the superstructure on a highway bridge with high precision without considering the positions of joints.

Maintenance of elevated highway bridges built during the high economic growth period is performed based on their design drawings and as-built drawings. For these elevated highway bridges, however, drawings were produced on paper media at the times of design and completion; therefore, in many cases they have already been disposed, it is difficult to make it for performing maintenance. In such a case, regenerating the detail design of the current status is required, but it takes huge cost. For this problem, a large number of studies have been made on automatic generation of three-dimensional models for maintenance of highway bridges from point cloud data obtained through MMS (Mobile Mapping System). However, it is hard to use the models generated in these studies because they do not consider the original geometric information of highways. Thus in this study, we aim to regenerate the SXF drawings of the superstructure of an elevated highway bridge, and propose a method for extracting alignment information of the superstructure from point cloud data of MMS. We verify the usability of our method and system by comparing the experimental data with current drawings in our experiments.

The elevated highway bridges intensively developed during the high economic growth period are now becoming deteriorated over time all together. This requires a maintenance technology to preserve highway functions systematically. To make a systematic maintenance plan, it is essential to grasp the present condition of road surface. This requires a route survey by taking some measures such as temporary suspension of traffic, though it is not easy. This study proposes a method for creating three-dimensional (3D) current state drawings that keep information such as road centerlines and are usable for maintenance on the superstructure of highway bridge by performing a route survey with a mobile mapping system (MMS). Our demonstration experiments verify the validity in making a maintenance plan with current state drawings in comparison with the existing materials such as final drawings of construction works.

Recently, an electronic delivery for various documents is carried out by Ministry of Land, Infrastructure, Transport and Tourism in construction fields. One of them is image data of construction photography that must be delivered with information of photography management items such as construction name or type of works, etc. However, there is a problem that a lot of cost is needed to treat contents of these items from characters printed and handwritten on blackboard into these image data. In this research, we develop the system which can treat contents of these items by extracting contents of these items from the image data of construction photography taken in various scenes with preprocessing the image, recognizing characters with OCR and correcting error with natural language process. And we confirm the effectiveness of the system, by experimenting in each function of system and in entire system.