丸山 喜久

Lifelines such as drinking water and sewage systems provide the means and conveyance for daily critical services, and they are essential systems for recovery operations after a damaging earthquake. Therefore, earthquake damage to lifeline components needs to be reliably assessed for a possible future seismic scenario. This study presents an earthquake damage assessment of buried pipeline networks in the Lima Metropolitan Area (LMA). It includes seismic hazard analysis, a review of pipeline network datasets, and the selection of empirical fragility functions. Deterministic seismic hazard analysis was performed for an inter-plate earthquake scenario using ground motion prediction equations and site conditions to compute the distribution of the peak ground velocity (PGV). Recommenda-tions are offered for an adequate selection of fragility functions developed in other regions, and a logic tree of fragility functions is proposed to be used in pipelines of LMA according to the data of pipeline damage after the 2007 Pisco Earthquake. Finally, the pipeline repair ratios and the total number of repairs are estimated for the earthquake scenario, and the results are geographically presented for each pipeline network.

The restoration period of the water supply system in Lima, Peru, after a scenario earthquake was estimated in this study. To achieve the objective, the probabilistic assessment model for post-earthquake residual capacity of the utility lifeline system initially proposed by Nojima and Sugito (2005) and revised by following related studies was employed. The dataset of water distribution pipelines was provided by Potable Water and Sewer System Service in Lima, Peru (SEDAPAL), and the spatial distribution of ground motion with a moment magnitude of 8.6 was considered as a scenario earthquake in this study. The water disruption was anticipated to continue for approximately one month in certain districts of Lima, Peru. The estimated smallest water supplying ratio was 21.1% in Villa El Salvador after the scenario earthquake.

2016年4月熊本地震では，熊本県熊本地方を震源とするMw6.2の地震が発生し，その約16時間後に同地域を震源とするMw7.0の地震が発生した．本研究では，地震による建物の被害状況を効率的にかつ安全に把握する方法として，航空レーザ測量データを深層学習することによって建物被害検出モデルの構築を試みた．本震前後に収集した航空レーザ測量データに対し，深層学習のアルゴリズムの一つである畳み込みニューラルネットワーク（CNN）を適用し，ネットワーク構成を変えながら最良のモデルの検討を行った．その結果，正答率が90%を超えるモデルを構築することができた．

現在，日本の水道では年間2万件を超える漏水・破損事故が発生している．上水道管の漏水は，地上に流れ出す地上漏水と，地上には流れ出さず地下で流れている地下漏水の2種類に大別できる．地上漏水は人目に触れることから発見しやすいものの，地下漏水は漏水の状況を直接目視で確認できないため，早期発見のための技術開発が求められている．そこで本研究では，現在普及が進んでいるスマートメータを活用した水道管路のモニタリングを想定し，管網端部の水圧情報を使用した漏水位置予測に関する検討を行った．漏水シナリオや機械学習手法の異なる6つのモデルを構築し，その予測精度を比較した．水圧変化率，水圧変化量，管種情報を説明変数とし，LightGBMに基づき構築した漏水予測モデルが最も良好な結果を示した．

Road markings, including road lanes and symbolic road markings, can convey abundant guidance information to autonomous driving cars. However, recent works have paid less attention to the recognition of symbolic road markings compared with road lanes. In this study, a road-marking-segmentation dataset named the RMD (Road Marking Dataset) is introduced to compensate for the lack of datasets and the limitations of the existing datasets. Furthermore, we propose a novel multiscale attention-based dilated convolutional neural network (MSA-DCNN) to tackle the proposed RMD. The proposed method employs multiscale attention to merge the weighting outputs of adjacent multiscale inputs, and dilated convolution to capture spatial-context information. The performance analysis shows that the proposed MSA-DCNN yields the best results by combining multiscale attention and dilated convolution. Additionally, the proposed method gains the mIoU of 74.88%, which is a significant improvement over the existing techniques.

Remote sensing is an effective method of evaluating building damage after a large-scale natural disaster, such as an earthquake or a typhoon. In recent years, with the development of computer vision technology, deep learning algorithms have been used for damage assessment from aerial images. In April 2016, a series of earthquakes hit the Kyushu region, Japan, and caused severe damage in the Kumamoto and Oita Prefectures. Numerous buildings collapsed because of the strong and continuous shaking. In this study, a deep learning model called Mask R-CNN was modified to extract residential buildings and estimate their damage levels from post-event aerial images. Our Mask R-CNN model employs an improved feature pyramid network and online hard example mining. Furthermore, a non-maximum suppression algorithm across multiple classes was also applied to improve prediction. The aerial images captured on 29 April 2016 (two weeks after the main shock) in Mashiki Town, Kumamoto Prefecture, were used as the training and test sets. Compared with the field survey results, our model achieved approximately 95% accuracy for building extraction and over 92% accuracy for the detection of severely damaged buildings. The overall classification accuracy for the four damage classes was approximately 88%, demonstrating acceptable performance.

平成30年7月豪雨は，西日本を中心に広範囲に甚大な被害をもたらした．このような豪雨の発生は増加傾向であり，土砂崩壊発生地点を予測することは重要である．現在行われている土砂災害警戒区域等の指定は，労力と時間が非常にかかるうえ，指定状況に関する課題もある．そこで本研究では，機械学習の一手法であるランダムフォレストを用いて，土砂崩壊発生地点の予測モデルを構築した．説明変数の異なる2つのモデルを作成し，その予測精度を比較した．また，近年研究が進められている説明可能AI（XAI）の一つであるSHAPを用いて大域説明，局所説明を行い，予測結果の判断根拠を明らかにした上で2つのモデルの結果に差異が生じた理由の考察を行った．

2018年北海道胆振東部地震では，厚真町やその周辺で斜面崩壊が多発し，多くの犠牲者が発生した．既往研究では，斜面崩壊予測方法が提案され推定精度に関する検討が行われてきたが，広域に適用可能な予測方法は未だ研究段階である．そこで本研究では，機械学習を用いて北海道胆振東部地震における斜面崩壊の推定を行い，その精度を検証した．さらに，衛星画像が地震後早期に取得できたものと仮定し，正規化植生指数（NDVI）を機械学習の説明変数に用いることを検討した．推定結果をκ係数等に基づき評価すると，ランダムフォレストによる推定結果が最も精度が高く，特にNDVIが利用できる場合は推定結果と実際の斜面崩壊の一致度が良好であった．

東北地方太平洋沖地震では，東日本の広域にわたって液状化が発生した．現在の液状化予測では詳細な地盤調査データを利用した手法がよく用いられているが，このような手法を広域に適用することは困難である．そのため，道路や水道，都市ガスのようなライフラインネットワークの被害予測には，簡便に面的な評価を行える手法の確立が望まれる．そこで，本研究ではサポートベクターマシンとランダムフォレストを用いて，液状化発生予測モデルを構築することを目的とする．既往研究の検討結果を踏まえて，液状化の発生に影響がある地震動継続時間を変数として新たに加えた．また，液状化の発生傾向によってモデルを微地形区分に基づきグループ分けすることを検討した．

Typhoon Hagibis passed through Japan on October 12, 2019, brought heavy rainfalls over half of Japan, and resulted in flooding in wide areas. This study focuses on a flooded state-managed river, the Naka River, Ibaraki Prefecture, Japan. In this study, we conducted the field survey on October 28, 2019, after the typhoon passed. Additionally, a series of runoff analyses for the entire Naka River basin and inundation analyses for some sections of the Naka River were conducted. Comparing the results of numerical simulation with the inundated areas detected by aerial and satellite Synthetic Aperture Radar (SAR) images, the accuracy of the numerical simulation was investigated. The maximum F value was 78.65%, and the simulated result broadly coincided with the inundated areas.

Bridges are an important part of road networks in an emergency period, as well as in ordinary times. Bridge collapses have occurred as a result of many recent disasters. Synthetic aperture radar (SAR), which can acquire images under any weather or sunlight conditions, has been shown to be effective in assessing the damage situation of structures in the emergency response phase. We investigate the backscattering characteristics of washed-away or collapsed bridges from the multi-temporal high-resolution SAR intensity imagery introduced in our previous studies. In this study, we address the challenge of building a model to identify collapsed bridges using five change features obtained from multi-temporal SAR intensity images. Forty-four bridges affected by the 2011 Tohoku-oki earthquake, in Japan, and forty-four bridges affected by the 2020 July floods, also in Japan, including a total of 21 collapsed bridges, were divided into training, test, and validation sets. Twelve models were trained, using different numbers of features as input in random forest and logistic regression methods. Comparing the accuracies of the validation sets, the random forest model trained with the two mixed events using all the features showed the highest capability to extract collapsed bridges. After improvement by introducing an oversampling technique, the F-score for collapsed bridges was 0.87 and the kappa coefficient was 0.82, showing highly accurate agreement.

The extraction of roads and centerlines from aerial imagery is considered an important topic because it contributes to different fields, such as urban planning, transportation engineering, and disaster mitigation. Many researchers have studied this topic as a two-separated task that affects the quality of extracted roads and centerlines because of the correlation between these two tasks. Accurate road extraction enhances accurate centerline extraction if these two tasks are processed simultaneously. This study proposes a multitask learning scheme using a gated deep convolutional neural network (DCNN) to extract roads and centerlines simultaneously. The DCNN is composed of one encoder and two decoders implemented on the U-Net backbone. The decoders are assigned to extract roads and centerlines from low-resolution feature maps. Before extraction, the images are processed within an encoder to extract the spatial information from a complex, high-resolution image. The encoder consists of the residual blocks (Res-Block) connected to a bridge represented by a Res-Block, and the bridge connects the two identical decoders, which consists of stacking convolutional layers (Conv.layer). Attention gates (AGs) are added to our model to enhance the selection process for the true pixels that represent road or centerline classes. Our model is trained on a dataset of high-resolution aerial images, which is open to the public. The model succeeds in efficiently extracting roads and centerlines compared with other multitask learning models.

Typhoon Hagibis passed through Japan on October 12, 2019, bringing heavy rainfall over half of Japan. Twelve banks of seven state-managed rivers collapsed, flooding a wide area. Quick and accurate damage proximity maps are helpful for emergency responses and relief activities after such disasters. In this study, we propose a quick analysis procedure to estimate inundations due to Typhoon Hagibis using multi-temporal Sentinel-1 SAR intensity images. The study area was Ibaraki Prefecture, Japan, including two flooded state-managed rivers, Naka and Kuji. First, the completely flooded areas were detected by two traditional methods, the change detection and the thresholding methods. By comparing the results in a part of the affected area with our field survey, the change detection was adopted due to its higher recall accuracy. Then, a new index combining the average value and the standard deviation of the differences was proposed for extracting partially flooded built-up areas. Finally, inundation maps were created by merging the completely and partially flooded areas. The final inundation map was evaluated via comparison with the flooding boundary produced by the Geospatial Information Authority (GSI) and the Ministry of Land, Infrastructure, Transport, and Tourism (MLIT) of Japan. As a result, 74% of the inundated areas were able to be identified successfully using the proposed quick procedure.

Record-breaking heavy rainfall hit Japan from July 3 to 31, 2020. The water level of rivers rose rapidly, and many bridges were washed away. In this study, two temporal ALOS-2 PALSAR-2 intensity images were introduced to detect damaged bridges over the Kuma River in Yatsushiro City and Ashikita Town, Kumamoto Prefecture, Japan. First, the backscattering models of bridges were investigated. Then the damage conditions of twenty-five target bridges were examined by the difference and correlation coefficient in the backscatter intensity. In addition, the increases of water levels was estimated by the movements of bridges’ SAR model. A field survey report and a post-event SPOT-7 satellite optical image were used to verify our results.

This paper investigates the 2016 Kumamoto Earthquake to explore the use of post-event full polarimetric SAR images in building damage assessment. In this study, one post-event airborne SAR image of Pi-SAR-X2 was employed to detect the changes of residential buildings in Mashiki Town, Kumamoto Prefecture, Japan. The changes in scattering characteristics of collapsed and non-collapsed buildings were compared using variance, homogeneity, contrast and dissimilarity textures of the Grey Level Co-occurrence Matrix (GLCM) from building footprints and roof areas. In addition, due to the difficulty in detection using only one texture, combination of textures in different polarizations were also applied for damage detection. The results of this method were then compared with the field survey data. The result indicates the possibility of building damage assessment using only post-event data and solving the pre-event data limitation problem.

<p>本研究では，地震により被災した建物を対象に実施される住家の被害認定調査の効率化を目的とし，深層学習のアルゴリズムの一つである畳み込みニューラルネットワーク（CNN）を適用した建物の被災度判別を試みた．まず，2016年熊本地震後に実施された被害認定調査における建物の外観画像からデータセットを作成した．そのデータセットにCNNを適用し，対象建物の全壊・非全壊を判別するモデルの構築に向けた検討を進めた．画像数の不足に伴う過学習を抑制するために，他の地震で被災した建物の画像データを追加し，判別モデルの精度向上を図ることができた．</p>

2019年9月に関東地方に上陸した令和元年房総半島台風（台風第15号）（TY1915）では，強風と豪雨により東京電力管内の2本の鉄塔と1,996本の電柱の破損，倒壊などの被害が発生した．台風の影響により，千葉県，静岡県，山梨県，神奈川県，茨城県，群馬県，栃木県，東京都，埼玉県の合計で約93万軒の停電が発生した．復旧が困難だった一部の地域を除いても，停電の復旧には約2週間を要したため，電気通信や水道などの他のライフラインに加えて社会・経済活動に大きな影響が発生した．本報告書では，TY1915に関する経済産業省の公開データ，東京電力配電網の電柱被害データ，東京電力パワーグリッドの停電履歴の公開データに加え，災害時情報集約支援チームISUTによる倒木データ，独自に実施した風況再現結果を用いて送配電設備の被害と停電発生の状況に関する定量的な分析結果を取りまとめた．In Typhoon Faxai (TY1915) that landed on the Kanto region in September 2019, strong winds and heavy rain caused damage such as breakage and collapse of two steel towers and 1,996 power poles in the jurisdiction of TEPCO. The impact of the typhoon has led to a total of about 930,000 households in Chiba, Shizuoka, Yamanashi, Kanagawa, Ibaraki, Gunma, Tochigi, Tokyo and Saitama prefectures. The restoration of the power outage took about two weeks, except in some areas where restoration was difficult, affecting other lifelines such as telecommunications and water supply. In this report, the results of surveys on damage to power transmission and distribution facilities and power outages caused by TY1915 were compiled using public data from the Ministry of Economy, Trade and Industry, power pole damage data, power outage history data from TEPCO Power Grid Co. Ltd and fallen tree data aggregated by Information Support Team.(ONLINE先行公開につき、ページ数および発行年情報は仮のものになります。引用にはDOIをご利用ください)

2019年度はデータ利活用協議会におけるデータ利活用のためのルールと規格を検討し、分科会活動を進めた。地震波形情報等を保有する企業と個別に「データ提供に関する覚書」を結び、民間企業のデータを活用する準備を行った。当初は3年目中までに4社との契約を目指したが、すでに2年目中に4社(令和元年度現在)と契約が結べ、防災科学技術研究所の基盤的データ、首都圏地震観測網データの統合が進んだ。 新潟大が中心となり、データ利活用協議会の活動を推進した。参加者数は1,165名となった。組織会員は、2018年度より10増えて67会員（企業54、自治体は3、団体は9、その他1団体）、個人会員は13名となった。データ利活用協議会では、理事会による協議会の運営体制が確立した。分科会活動についても積極的に推進した。具体的には、7分科会と1ワーキンググループにおいて、合計で年間で48回（内訳：会議・打合せ38回、調査9回、イベント1回）を実施した。また、分科会活動全体のデータ利活用を統合的に推進するインフラ分科会を新たに立ち上げた。 防災科研が中心となり、防災科学技術研究所が保有している情報インフラ基盤に、首都圏レジリエンスプロジェクトとPRISMから得られるデータの連携や2次利用に関して、技術的課題を抽出し、対応策を検討した。特に、企業とのデータ連携については、防災情報サービスプラットフォームを介した情報の共有について、その流れや権利関係等の課題も含め検討した。 東京工業大が中心となって、被害拡大阻止のために、2016熊本地震の宇城市の被害認定調査結果および地震動分布を整理し、宇城市の建物フラジリティ関数を構築した。また、地震動レベルの大きい益城町とやや小さい宇城市の建物フラジリティを統合する方法について検討した。また、益城町の現地写真に基づく深層学習や写真目視により層破壊建物データを作成し、その建物の分布を明らかにすると共に、航空レーザーデータを用いた層破壊建物の抽出手法を検討した。さらに2018北海道胆振東部地震の地震動データの収集と整理を行い、被災地の地震動分布を推定すると共に安平町の罹災証明データの収集と分析を進めた。 岐阜大が中心となって、供給・処理・交通系ライフラインの被害・復旧データを収集・整理し、地震動分布との統計分析により、物理的・機能的フラジリティ関数、レジリエンス関数を構築した。１）上水道管路は2007新潟県中越沖、2011東北地方太平洋沖、2016熊本地震時の上水道管路の被害データを用い、最大速度（PGV）、管種、液状化の影響度が大きいことを明確化、２）下水道埋設管路は、2018北海道胆振東部地震で被災した下水道埋設管路網のデータベース化と被害率算定の実施、3）鉄道に関しては、2011東北地方太平洋沖、2016熊本、2018大阪府北部、2018北海道胆振東部地震における機能的フラジリティ関数を構築、３）デ活・インフラ分科会の立上げに貢献、等を行った。 富山大が中心となって、1）2019山形沖地震の早期被害把握を目指してAIによるブルーシートの判定を実施、2）2019台風15号では同判定モデルを活用・検証を実施（35%の適合率に留まる）、3）既述1）2）により、ドローンの活用には地域特性を含めたAIの学習機会を創出する仕組みが必要であることを明確化、4）2018大阪北部の地震においては衛星画像を用い、深層学習・ニューラルネットワークの1つであるVGG16モデルを用い、短時間の学習で、適合率93.75％、再現率94.74％（F値：0.94）を実現した。これらの結果から、衛星画像とドローンによる収集画像を組み合わせることで、精度の向上が見込まれるという見通しを実証した。 兵庫県立大が中心となって、事業継続能力向上のための業務手順の確立を目指し、１）2019台風第15号・第19号・第21号における企業の事業継続活動に関する質問紙調査を行い、BCP（事業継続計画）が機能しなかった理由等を明確化、２）災害対応に工程（プロセス）の視点を持ちこんだ災害対応工程管理システムBOSSにおいては、2019台風19号被災地・静岡県南伊豆町を事例として自治体職員実装による検証・改良を実施、３）「帰宅困難者支援施設運営ゲーム（企業内待機版）KUG）について、実際の帰宅困難者支援施設運営ゲーム（KUG②）ワークショップによる複数回の検証を経て（東京都台東区、東京都渋谷区、京都府、自治体主催）津波なしシナリオゲームを新たに作成した。 関西大学が中心となって、1）南海トラフ巨大地震「事前避難」防災計画策定について、「津波避難対策特別強化地域」半数の未策定自治体に対し、新たな文化的防災を提案する等方向性を提示、2）首都直下地震おいては、縮災対策の内容としてネットワーク社会のノードに対する複合災害とエッジに対する連続滝状災害を阻止することが重要であることを明確化、3）2019九州地方の水害・台風15号・19号災害において、調査を実施、府県レベルの地域防災計画の改定内容を提示すると共に台風19号のコースが東寄りで、東京に集中豪雨が発生すれば荒川などの首都圏の河川が氾濫したことを実証・対策の必要性を明確化、した。

S: Remote sensing is an effective method to evaluate the damage situation after a large-scale nature disaster. Recently, deep learning algorithms have been used for the damage assessment from remote sensing images. A series of earthquakes hit the Kyushu region, Japan in April 2016, and caused severe damage in Kumamoto and Oita Prefectures. Numerous buildings were collapsed by the continuous strong shaking. In this study, the authors modified the Mask R-CNN model to extract residential buildings and estimate their damage levels. The Mask R-CNN model employs a two-stage instance segmentation algorithm which maintains a Convolutional Neural Network backbone and a Region Proposal Network with a ROI Align head. The aerial images captured on April 29, 2016 (two weeks after the main shock) in Mashiki Town, Kumamoto Prefecture, were used as the training and test sets. Comparing with the damage report of the field survey, the accuracy for the building extraction was 92%. As for the damage estimation, the precision and recall of the collapsed buildings achieved approximately 72% and 95%.

<p> 本研究では地理情報システム(GIS)を用いて，2016年熊本地震における上水道管路の被害分析を行った．上水道管路の管種，継手，口径，微地区分，地震動強さと被害率の関係性を分析し，現行の被害予測式の補正係数と比較した．また，液状化や微地形の境界条件を考慮して被害分析を行った．基準属性の管路被害率に対する比と既往の被害予測式の口径補正係数，微地形補正係数は近い値となったが，管種・継手補正係数に関しては一致しない場合もあった．そこで，近年の地震時の上水道管路の被害データに基づき，管路属性，地震動強さ，液状化，微地形の境界条件が管路被害率に与える影響度を明らかにするため，数量化理論I類による回帰分析を行った．</p>

<p> 本研究の目的は，H/Vスペクトル比の機械学習に基づき地盤増幅度を推定することである．そこで，本研究では，地盤増幅度と相関が高い地盤特性情報である深さ30mまでの地盤の平均S波速度（AVS30）を地震動H/Vスペクトル比の機械学習に基づき推定した．はじめに，日本全国に配備されているK-NETおよびKiK-netの観測データを使用し，常時微動H/Vスペクトル比と振幅形状が似ているとされる地震動H/Vスペクトル比を計算する．次に，公開されている観測点のS波速度構造からAVS30を算出し，両者の関係性を考察した．地盤増幅度の評価のために，機械学習に基づく回帰分析によって地震動H/Vスペクトル比からAVS30を推定することを試みた．さらに，深部地盤情報データや微地形区分を取り入れることで，本研究の予測モデルの精度向上を図った．</p>

<p> 東京ガス（株）は地震時の2次災害防止を目的として，供給区域内約4,000箇所に設置した超高密度な地震計から構成されるリアルタイム地震防災システム"SUPREME"を導入している．SUPREMEは地震観測情報を速やかに収集し，即時に供給停止すべきブロックを特定するとともに低圧ガス導管の被害予測を行い，被害程度に応じた供給停止判断を支援する重要な役割を担っている．本研究では，2011年に発生した東日本大震災（東北地方太平洋沖地震）における低圧ガス導管の被害に対して，被害形態や地形影響を考慮した詳細な分析・評価を行った．最終的には，ねじ接合鋼管の被害を継手と管体に区分し，地形分類ごとに独立した被害予測式を作成することで被害予測システムの高精度化を図った．</p>

<p> 令和元年房総半島台風と命名された2019年台風15号は，非常に強い勢力を保ったまま9月9日に千葉県千葉市に上陸した．千葉県内では7万棟以上の住家が一部損壊以上の被害を受けた．そこで本研究では，国際航業（株）と朝日航洋（株）がそれぞれ9月19～20日，9月27～28日に撮影した航空写真を用いて，千葉県房総半島の建物の屋根損傷部に覆われたブルーシートを抽出した．一部地域で目視判読を行い，それを検証用データとしてブルーシートの自動抽出方法を提案した．なお，国際航業と朝日航洋が撮影した航空写真の撮影条件が異なるため，両者には異なる自動抽出方法を使用した．最後に，ブルーシートの面積と屋根投影面積によって算出したブルーシート被覆率を最大瞬間風速と比較し，最大瞬間風速による屋根のブルーシート被覆率の予測関数を構築した．</p>

Updating road networks using remote sensing imagery is among the most important topics in city planning, traffic management and disaster management. As a good alternative to manual methods, which are considered to be expensive and time consuming, deep learning techniques provide great improvements in these regards. One of these techniques is the use of deep convolution neural networks (DCNNs). This study presents a road segmentation model consisting of a skip connection of U-net and residual blocks (ResBlocks) in the encoding part and convolution layers (Conv. layer) in the decoding part. Although the model uses fewer residual blocks in the encoding part and fewer convolution layers in the decoding part, it produces better image predictions in comparison with other state-of-the-art models. This model automatically and efficiently extracts road networks from high-resolution aerial imagery in an unexpansive manner using a small training dataset.

© 2019 by the authors. A series of earthquakes hit Kumamoto Prefecture, Japan, continuously over a period of two days in April 2016. The earthquakes caused many landslides and numerous surface ruptures. In this study, two sets of the pre- and post-event airborne Lidar data were applied to detect landslides along the Futagawa fault. First, the horizontal displacements caused by the crustal displacements were removed by a subpixel registration. Then, the vertical displacements were calculated by averaging the vertical differences in 100-m grids. The erosions and depositions in the corrected vertical differences were extracted using the thresholding method. Slope information was applied to remove the vertical differences caused by collapsed buildings. Then, the linked depositions were identified from the erosions according to the aspect information. Finally, the erosion and its linked deposition were identified as a landslide. The results were verified using truth data from field surveys and image interpretation. Both the pair of digital surface models acquired over a short period and the pair of digital terrain models acquired over a 10-year period showed good potential for detecting 70% of landslides.

© 2019 by the authors. Timely information about landslides during or immediately after an event is an invaluable source for emergency response andmanagement. Using an active sensor, synthetic aperture radar (SAR) can capture images of the earth's surface regardless ofweather conditions andmay provide a solution to the problem of mapping landslides when clouds obstruct optical imaging. The 2018 Hokkaido Eastern Iburi earthquake (Mw 6.6) and its aftershocks not only caused major damage with severe loss of life and property but also induced many landslides across the area. To gain a better understanding of the landslides induced by this earthquake, we proposed a method of landslide mapping using pre- and post-event Advanced Land Observation Satellite 2 Phased Array L-band Synthetic Aperture Radar 2 (ALOS-2 PALSAR-2) images acquired from both descending and ascending orbits. Moreover, the accuracy of the classification results was verified by comparisons with high-resolution optical images, and ground truth data (provided by GSI, Japan). The detected landslides show a good match with the reference optical images by visual comparison. The quantitative comparison results showed that a combination of the descending and ascending intensity-based landslide classification had the best accuracy with an overall accuracy and kappa coefficient of 80.1% and 0.45, respectively.

© 2019 IEEE. Successive heavy rainfall affected the western Japan from the late June to the early July 2018. Increased river water overflowed and destroyed river banks, which caused flooding in vast areas. In this study, two pre-event and one co-event ALOS-2 PALSAR-2 images were used to extract inundation areas in Kurashiki and Okayama Cities, Okayama Prefecture, Japan. First, the difference between the pre-event and co-event coherence values was calculated. The decreased coherence areas were extracted as possible inundation. Then the water regions were extracted by the threshold values from the threeoral intensity images. The increased water regions in July 2018 were obtained as inundation. Finally, the extracted results from the coherence and intensity images were merged to create an inundation map. The results were verified by comparing with a web-based questionnaire survey report and visual interpretation of aerial photos.

© 2019, Fuji Technology Press. All rights reserved. A series of heavy rainfalls hit the western half of Japan from June 28 to July 8, 2018. Increased river water overflowed and destroyed river banks, causing flooding over vast areas. In this study, two pre-event and one co-event ALOS-2 PALSAR-2 images were used to extract inundation areas in Kurashiki and Okayama Cities, Okayama Prefecture, Japan. First, water regions were extracted by threshold values from three-temporal intensity images. The increased water regions in July 2018 were obtained as inundation. Inundated built-up areas were identified by the increase of backscattering intensity. Differences between the pre-and co-event coherence values were calculated. The area with decreased coherence was extracted as a possible inundation area. The results of a field survey conducted on July 16, 2018 were used to estimate the optimal parameters for the extraction. Finally, the results from the intensity and coherence images were verified by making comparisons between a web-based questionnaire survey report and the visual interpretation of aerial photographs.

<p>本研究では, 車載カメラから取得した画像から, 地震による道路被害を深層学習によって自動で抽出することを目的としている．あらかじめ目視で被害の有無を分類した教師用画像を使用し, それらを畳み込みニューラルネットワーク（CNN）で深層学習し, 画像判別モデルの作成を行った．この画像判別モデルに学習に使用していない精度評価用画像300枚を判別させたところ, 道路閉塞の判別精度は87%, 無被害の判別精度は90%と高かったが, 道路被害の判別精度は66%とやや低かった．この結果を踏まえて, 地震直後の利用を想定した条件設定を行い, 道路変状の自動抽出シミュレーションを行った．本研究の手法は地震後の道路被害の早期把握に有効と考えられ, 道路管理者の震後対応に貢献できる．</p>

<p>The Road Committee of the Panel on Infrastructure Development recommended full-scale maintenance of aging roads in 2014. According to their recommendations, pavements may be inspected or replaced based on an appropriate renewal period depending on their deterioration level. The authors proposed a method to evaluate road surface irregularity using vertical accelerations recorded by smartphone. This study applies the method to a daily road patrol performed by road administrator of local governnment. A smartphone is installed in an official car of Nogata City Government, Fukuoka Prefecture. The accelerations are recorded during the daily road patrol, and the road surface irregularity is evaluated on a routine schedule.</p>

In this study, the characteristics of site amplification at seismic observation stations in Japan were estimated using the attenuation relationship of each station's response spectrum. Ground motion records observed after 32 earthquakes were employed to construct the attenuation relationship. The station correction factor at each KiK-net station was compared to the transfer functions between the base rock and the surface. For each station, the plot of the station correction factor versus the period was similar in shape to the graphs of the transfer function (amplitude ratio versus period). Therefore, the station correction factors are effective for evaluating site amplifications considering the period of ground shaking. In addition, the station correction factors were evaluated with respect to the average shear wave velocities using a geographic information system (GIS) dataset. Lastly, the site amplifications for specific periods were estimated throughout Japan.

In exploring the relationship between ground-level road damage ratios and tsunami inundation depths following the 2011 Pacific Coast Tohoku earthquake in Japan, we focused on road damage components, excluding elevated roads, bridges, and tunnels. The damage ratio is defined as the number of damage incidents per kilometer. We used the damage dataset compiled by the Japanese Ministry of Land, Infrastructure and Transport. We propose four fragility function zones for ground-level roads based on differences in topographical features. We studied these zones based on numerical simulation results of tsunami propagation.

&lt;p&gt;本研究は, 東京湾北部地震における東京23区の木造建物の解体棟数を予測することを目的としている．まず, 神戸市企画調整局から提供された兵庫県南部地震後にまとめられた震災復興データアーカイブ内の建物データを用いて, 木造建物の地震時の解体損傷度を推定した．推定された兵庫県南部地震の際の解体損傷度と, 川口ら（2013）による新潟県中越地震時の解体損傷度を比較し, 解体損傷度の地震間の整合性に関して検討した．さらに, 東京湾北部地震を対象として, 東京23区の木造建物の解体棟数を予測した．&lt;/p&gt;

&lt;p&gt;熊本洋学校教師館ジェーンズ邸は県指定の重要文化財であり，2016年熊本地震の強震動の作用によって倒壊した．本研究では，倒壊地点における地震動を明らかにすることを目的とし，倒壊地点等において余震観測を実施した．得られた記録を分析したところ，倒壊地点に作用した地震動は，近傍のガバナ内に西部ガス (株) が設置している観測点 (水前寺ガバナ) で得られた前震・本震記録と同等程度であることが明らかとなった．さらに，サイト増幅特性置換手法に基づき倒壊地点における前震時・本震時の地震動を推定し，その際，水前寺ガバナでの観測記録を用いて地震動推定手法の適用性を検証した．&lt;/p&gt;

本研究では，スマートフォンで計測した自動車の上下加速度を用いて路面凹凸を評価する数理モデルを構築した．ロジスティック回帰分析とサポートベクトルマシンの2種類の機械学習手法に基づき，国際ラフネス指数（IRI）が12 mm/m以上の区間を抽出する数理モデルを構築することを目的とした．数理モデルの構築に使用していないデータを用いて，2種類の手法による数理モデルの判定精度を評価したところ，ロジスティック回帰分析の方が良好な結果を示した．さらに，路面不良区間が誤判定される原因として，スマートフォンで加速度を取得するときの車速が影響していることが分かった．

The 2011 Tohoku event showed the massive destruction potential of tsunamis. The Euro-Japan "Risk assessment and design of prevention structures for enhanced tsunami disaster resilience (RAPSODI)" project aimed at using data from the event to evaluate tsunami mitigation strategies and to validate a framework for a quantitative tsunami mortality risk analysis. Coastal structures and mitigation strategies against tsunamis in Europe and Japan are compared. Failure mechanisms of coastal protection structures exposed to tsunamis are discussed based on field data. Knowledge gaps on failure modes of different structures under different tsunami loading conditions are identified. Results of the wave-flume laboratory experiments on rubble mound breakwaters are used to assess their resilience against tsunami impact. For the risk analysis, high-resolution digital elevation data are applied for the inundation modeling. The hazard is represented by the maximum flow depth, the exposure is described by the location of the population, and the mortality is a function of flow depth and building vulnerability. A thorough search for appropriate data on the 2011 Tohoku tsunami was performed. The results of the 2011 Tohoku tsunami mortality hindcast for the city of Ishinomaki substantiate that the tsunami mortality risk model can help to identify high-mortality risk areas and the main risk drivers.