圖書典藏及影音數位平台

本局將於107~108年6月完成「頭前溪智慧河川建置計畫第一期」計畫(以下簡稱第一期)，並期望智慧河川整體計畫完成後，產製出有用之防災數據，達到智慧河川支援決策之需求及目標。 本局之第一期的工作目標為自動化，主要完成儀器監控、模式預警統計、觀測數據輔助驗證及系統整合展示等四個面向工作。其中完成四個預警系統(分別為頭前溪流域(含桃園市)內水淹水全時預報系統、頭前溪舊港島及中港溪米粉街外水溢淹全時預報系統、頭前溪舊港島及中港溪米粉街(含桃園市)重要防汛熱點淹水範圍研判系統及蜆仔溝閘門啟閉監測系統)，建立初步智慧河川目標中早期預警、現地自動示警、決策支援與展示介面等機制。 本(第二期)計畫係接續第一期計畫成果目標為自動化及智慧化，依本局轄管河川特性補強觀測儀器與洪水相關預警，其中觀測儀器重點在河床自動沖刷監測與堤防安全地震示警監測，以提供本局同仁於汛期或地震時研判資訊。另外，在預警系統中，擴充移動式抽水機布設、搶修險備料輔助決策、現地勘查資料即時展示作業、河防構造物外水溢淹監控及即時災損評估等。此外，本(第二期)計畫同步開發智慧化的前瞻技術，例如：人工智慧自動回話與聊天機器人或人工智慧語音辨識等，進行防災跨領域整合，俾使計畫成果能更趨近智慧河川目標。 最後，第一期開發之四個全時運作系統，完成迄今均尚未歷經較大颱洪事件，後續驗證與調整一併納入本(第二期)計畫檢討與修正。

The First period of the Forward-looking technology of the Touqian River river construction plan was completed in June 108. The work objective of the first phase is automation. The second phase plan is the succession of the first phase of the plan. The goal is automation and intelligence. There are six major work items. 1. Continuous inventory of observation facilities 2. Planning and construction of automatic scour monitoring and earthquake warning 3. Expansion of intelligent intelligence platform. (1) Intelligent decision-making for the deployment of mobile pumps. (2) Intelligent decision-making for emergency repair insurance and auxiliary materials. (3) Intelligent on-site investigation. (4) Intelligent expansion of water overflow monitoring outside river defense structures. (5) Intelligent expansion of real-time disaster assessment. (6) Establish an intelligent warning mechanism. (7) Build an artificial intelligence login query system. (8) Conduct feasibility assessment of artificial intelligence voice query mechanism 4. Application and maintenance of existing systems 5. Continue to build a public outreach area the above work expects to effectively control and increase flood control response efficiency through intelligent methods to reduce the impact of errors caused by human judgment. Taking the decision-making of pump layout as an example, through numerical terrain and drainage simulation, we can clearly know the relative low point of the key area and gravity flow drainage. If large-scale flooding occurs, this decision can be used to quickly withdraw water. In the field survey operation, in the past, it was necessary to prepare complicated paper data, but it was not easy to be modified. Therefore, by combining the cloud database with LINE ROBOT, the step of data preparation can be omitted. Based on the above results, the future flood control and disaster prevention can be upgraded to another new milestone.