1242145 (1) 玖、發明說明 【發明所屬之技術領域】 本發明係關於支援以避免水災爲目的之防洪活動的防 洪支援裝置及程式、以及防洪支援方法。 【先前技術】 下水道或河川之防洪活動時，係依據氣象局發表之警 告或警報等來決定人員之班體制，班則會對應下水道或河 川之狀況而趕至現場。 以往亦有人提出取得雨量資訊並利用於防災活動之防 災支援裝置。 例如，專利文獻1之防災資訊提供裝置係具有：用以 儲存_量資訊、震度資訊、及失真資訊等防災資訊之防災 資訊記憶部；利用者持有之配備瀏覽器的機器；以及連結 於接收利用者之存取之配備瀏覽器的機器，且會輸出防災 資訊之讀取要求的WWW連結部。其次，專利文獻1之防 災資訊提供裝置會接收來自WWW連結部之讀取要求，判 斷是否提供防災資訊記憶部之防災資訊，判斷爲可提供時 ，會經由WWW連結部將防災資訊提供給利用者。 例如，專利文獻2之防災資訊收集處理系統則會以災 害現場之地點資訊做爲關鍵字，取已登錄之災害發生預測 地點的圖像資訊。其次，專利文獻2之防災資訊收集處理 系統會結合該圖像資訊及雨量資訊並進行編輯，然後，將 經過編輯之防災資訊儲存於記錄裝置，並經由資訊提供部 -5- 1242145 (2) 刊載於網際網路伺服器上，廣泛地提供給利用者。 [專利文獻1 ] 曰本特開平1 1 - 1 1 0454號公報 [專利文獻2 ] 日本特開2001-337997號公報 【發明內容】 氣象局之警告或警報，有時並非針對防洪對象區域之 貝5只°因此’依據氣象局之警告或警報來決定人員之班等 時，已決定之班有時只需保持待命狀態而無需採取任何具 體防洪活動。 上述傳統之防災資訊提供裝置或防災資訊收集處理系 統，會將依據雨量資訊等擬定之防災資訊廣泛地提供給利 用者’並利用於區域居民之防災宣導活動上。 然而，傳統之防災資訊提供裝置或防災資訊收集處理 系統並不具備支援預先防範水道或河川所導致之水災等廣 義防洪活動之機能《 一般而言，防洪活動包括對區域居民之宣導活動、現 場之監視、例如推土等之現場保安、以及將居民從危險場 所遷往安全場所等各種活動。 然而，因爲傳統之防災資訊提供裝置或防災資訊收集 處理系統只會對利用者提供防災資訊，以避免水災爲目的 之對策及其他防洪活動無法有效推行， 因此，針對水災之傳統上的一般防洪活動，係以氣象 -6- (3) 1242145 協會等特定機關針對地區發佈之大雨警告、大雨警報、洪 水警告、以及洪水警報等做爲指標。其次，傳統上之一般 防洪活動，係依據大雨警告、大雨警報、洪水警告、以及 洪水警報之其中任一項發佈所代表的資訊、及過去經驗， 推測水災發生場所，再將必要人員配置於防洪之必要場所 〇 然而，只將氣象資訊當做指標之防洪活動時，因爲氣 象資訊之發佈係以廣泛區域範圍爲對象，若發生實際從事 防洪活動之場所未下大雨，則會出現將人員指派至無需防 洪活動之場所執行防洪活動等情形，而有實施浪費資源之 防洪活動的問題。 有鑑於上述情形·）本發明之目的係在提供一種防洪支 援裝置及程式、以及防洪支援方法，可以防止對無需防洪 活動之場所實施防洪活動，且可對需要防洪活動之場所實 施防洪活動，實現效率良好之防洪活動。 以下係針對用以實現本發明之具體手段進行說明。 本發明第1形態係具有：取得氣象資訊之手段；取得 防洪支援之必要支援資訊手段；以及依據氣象資訊及支援 資訊，判斷防洪活動之必要性，判斷需要採取防洪活動時 ’擬定含防洪活動上必要之人員配備資訊在內之防洪支援 資δΤΙ之支援處理手段；之防洪支援裝置。 本發明第2形態係具有：取得氣象資訊之手段；取得 水災相關補助資訊之手段；依據補助資訊，取得防洪支援 之必要支援資訊補助資訊處理手段；以及依據氣象資訊及 (4) 1242145 龜 支援資訊，判斷防洪活動之必要性，判斷需要採取防洪活 - 動時，擬定含防洪活動上必要之人員配備資訊在內之防洪 ： 支援資訊之支援處理手段；之防洪支援裝置。 ' 本發明第3形態係具有：取得含有用以判斷防洪之過 - 去資訊、及對應該過去資訊之危險度在內之分析對象資訊 的手段；取得用以判斷防洪之現時點資訊的手段；以及依 據針對分析對象資訊實施資料挖掘所得到之預測手段、及 現時點資訊，擬定代表現時點之危險度的防洪支援資訊之 手段；之防洪支援裝置。 鲁 本發明第4形態係具有：取得含有用以判斷防洪之過 去資訊、及對應該過去資訊之危險度在內之教師資料的手 段；取得用以判斷防洪之現時點資訊的手段；以及依據利 用教師資料實施學習之類神經網路、及現時點資訊，擬定 代表現時點之危險度的防洪支援資訊之手段；之防洪支援 裝置。 【實施方式】 φ 以下’參照圖面針對本發明之實施形態進行說明。又 - ’以下之說明中，同一要素會附與相同符號，並省略其說 - 明。 * (第1實施形態） _ 1圖係本實施形態之防洪支援裝置實例的方塊圖。 防洪支援裝置1 0 1具有氣象資訊取得部1、補助資訊 -8- 1242145 (5) 耳5得部2、補助資訊處理部3、資訊設定部 部5、顯示裝置6、以及輸入裝置7a、7b。 氣象資訊取得部1係用以取得氣象資訊 例如代表是否會發生大雨警告、大雨警報、 及洪水警報等之資訊。 例如，氣象資訊取得部 1係用以 HIMAWARI1 1接收氣象資訊。又，例如， 部1可以從政府之氣象資訊提供機關經由網 收氣象資訊。又，例如，氣象資訊取得部1 中心及契約機關等氣象資訊提供契約機關1 象資訊。 補助資訊取得部2用以取得以補助氣象 水災相關補助資訊。 例如，補助資訊係包含利用設置於下水 洪對象區域之雨量計1 4檢測到之降雨量資 在內。 又，例如，補助資訊係包含利用水位計 下水道幹線及河川等之水位資訊在內。 又，補助資訊亦可包含和水災相關之其 內。補助資訊可以爲降雨量資訊、下水道幹 、及河川之水位資訊等之資訊當中之其中任 包含2種以上在內者。 補助資訊處理部3係用以執行將利用補 $取得之補助資訊轉換成防洪活動支援上必 4、支援處理 。氣象資訊係 洪水警告、以 從氣象衛星 氣象資訊取得 丨際網路1 2接 可從氣象業務 3直接接收氣 資訊爲目的之 道及河川等防 訊（雨量資訊） 1 5檢測到之 他各種資訊在 線之水位資訊 一者，亦可以 助資訊取得部 要之支援資訊 -9- 1242145 (6) 的處理。 資訊設定部4會設定著用以擬定防洪支援資訊之必要 資訊。 支援處理部5會擬定防洪活動上之必要防洪支援資訊 〇 顯示裝置6會顯示支援處理部5擬定之防洪支援資訊 〇 輸入裝置7 a、7 b係用於例如用以擬定防洪支援資訊 之必要資訊的輸入、及各種指示等之輸入。 第2圖係A縣B地區當中之防洪對象區域AB 1的下 水道幹線1 6 (或河川）、以及補助資訊檢測系之雨量計1 4 及水位計1 5之設置場所的關係圖。 該第2圖之實例中，在防洪對象區域AB !之必要場 所設置2個雨量計1 4。又，沿著流過防洪對象區域A B 1 之下水道幹線1 6設置2個水位計1 5。 氣象資訊取得部1從例如氣象資訊提供契約機關1 3 接收到大雨警告、大雨警報、洪水警告、以及洪水警報等 之氣象資訊時’會將接收到之氣象資訊提供給支援處理部 5 ° 補助資訊取得部2會接收例如含利用雨量計！ 4檢測 到之降雨量資訊、利用水位計1 5檢測到之水位資訊、及 其他水災（水災災害）相關資訊在內之補助資訊，並爲了利 用補助資訊擬定支援資訊，會將補助資訊、及檢測補助資 訊之機窃的設置場所辨識資料提供給補助資訊處理部3。 -10- (7) 1242145 補助資訊處理部3在例如補助資訊爲含有水位資訊在 內時，會針對各設置場所辨識資料，將補助資訊依各特定 時間（例如每5分鐘）轉換成代表水位變化之水位狀況資訊 ，並將含依時間系列得到之各特定時間的水位資訊及水位 狀況資訊在內之支援資訊記錄於資訊記錄部8。 以如上所示之方法來處理水位資訊時，會以設置場所 辨識資料做爲關鍵字，將含有水位資訊、水位狀況資訊、 及其他資訊在內之支援資訊記錄於資訊記錄部8。 又，補助資訊處理部3在例如補助資訊含有降雨量資 訊在內時，會針對各設置場所辨識資料，將補助資訊轉換 成開始下雨之開始累計之累計降雨量資訊、或各特定時間 之最大雨量資訊，並將依時間系列將含所得到之·累計降 雨量資訊、或各特定時間之最大降雨量資訊在內之支援資 訊記錄於資訊記錄部8。 以如上所示之方法處理降雨量資訊時，會以設置場所 辨識資料做爲關鍵字，將含有降雨量資訊、累計降雨量資 訊、最大降雨量資訊、及其他資訊在內之支援資訊記錄於 資訊記錄部8。 資訊設定部4設定著用以擬定如第3圖所示之防洪支 援資訊的必要資訊4 a。 亦即，資訊設定部4上針對各設置場所辨識資料「〇 ! 」、「〇2」、…設定著氣象資訊種別資料（D1 :大雨警告 、D2 :大雨警報、D3 :洪水警告、D4 :洪水警報）、閥値 、訊息、人員配置場所、以及各配備場所之配備人數等資 -11 - (8) 1242145 料。 又，閥値可針對降雨量資訊[mm/h]、累計降雨量資訊 [111111 ]、最大降雨量資訊[m m / h ]、水位資訊、水位狀況資 訊、以及其他資訊之其中任一種資訊來設定，然而，應針 對對下水道或河川而言爲最易受水災影響之資訊來進行設 定。 例如，亦可將降雨量分成「5 0 m m / h〜9 9 m m / h」、「 1 Ο 0 m m / h〜1 4 9 m m / h」、「1 5 0 m m / h以上」，並針對其中 一種資訊設定複數閥値。又，例如，亦可針對累計降雨量 資訊及水位資訊等2種以上之資訊設定閥値。 閥値亦可依據過去發生水災時之降雨量資訊、過去發 生水災時之下水道幹線的水位資訊、以及過去發生. 7_K災時 之河川的水位資訊等來決定。 例如，過去在50mm/h之降雨時發生過水災時，亦可 將5 Omm/h設定成降雨量之閥値。然而，因需對應降雨量 之增大來增加配備人數，故最好應針對複數降雨量設定閥 値，並對應各閥値設定配備人數。 又，過去未發生水災等時，亦可依據施設設計時計算 之計畫降雨量來設定閥値。 設定於資訊設定部4之訊息，因爲各下水道幹線或河 川、各設置場所辨識資料「0 1」、「02」、…並不相同， 可記述對應各下水道幹線或河川、及設置場所辨識資料之 內容。 因爲下水道幹線或河川之周圍往往會存在1個以上之 -12- (9) 1242145 危險場所’故可針對各下水道幹線或河川有效地設定複數 人員配置場所。 配備人數係針對各人員配置場所進行設定。 又，地區自治體有時會只針對1條河川之特定場所實 施防洪。此時，資訊4a中亦可省略設置場所辨識資料。 若不論任何氣象資訊種別資料皆實施防洪活動時，資 訊4a中亦可省略氣象資訊種別資料。 亦即，資訊設定部4上只需針對設置場所辨識資料、 閥値、及訊息之其中任一項，設定人員配置場所及各人員 配置場所之配備人數當中之其中任一資訊即可。 又，例如，依據降雨量資訊、下水道幹線之水位資訊 、及河川之水位資訊設定閥値時，亦可依據以過去統計資 料計算來設定閥値。 第4圖係過去之統計資料的實例圖。 係支援處理部5上記錄於資訊記錄部8之過去降雨量 資訊[m m / h ]或過去累計降雨量資訊[m m ]、下水道幹線之 過去水位資訊或河川之過去水位資訊的關係圖。又，可考 慮第4圖所示之過去水災狀況來決定閥値。 第4圖中，累計降雨量達到2 9 [ m m ]時，利用過去統 計資料之近似曲線的上限爲達到下水道幹線之警戒水位 21之0·6。因此，例如將閥値設定成累計降雨量29mm及 水位0.6m，並在累計降雨量或水位超過該閥値時執行防 洪活動之人員配置。 支援處理部5亦可自己執行邏輯演算、實施判斷，並 -13- (10) 1242145 輸出防洪支援資訊。又，支援處理部若由例如cpu所構 成時，亦可利用防洪處理程式執行特定處理並擬定防洪支 援資訊。 支援處理部5具有機能性之氣象判斷部5 1、必要性 判斷部5 2、擬定部5 3、及輸出部5 4。 氣象判斷部5 1會判斷氣象資訊取得部1接收之氣象 資訊是否爲防洪對象之氣象資訊。將例如大雨警告、大雨 警報、洪水警告、及洪水警報等氣象資訊定義成防洪對象 之氣象資訊。 必要性判斷部5 2在利用氣象判斷部5 1判斷爲防洪對 象之氣象資訊時，會參照資訊記錄部8之最接近現在時刻 之·例如降雨量資訊等支援資訊，該支援資訊若超過設定於 資訊設定部4之閥値時，會判斷必須實施防洪活動。 擬定部5 3在必要性判斷部5 2判斷需要採取防洪活動 時’會從資訊設定部4讀取訊息 '人員配置場所、及配備 人數資料等資訊’並以必要之格式實施編輯處理，擬定防 洪支援資訊。例如’防洪支援資訊亦可以爲畫面顯示用資 訊。 輸出部54會將擬定部53擬定之防洪支援資訊輸出至 顯示裝置6。 首先，輸入裝置7a'7b會輸入用以擬定防洪支援資 訊之必要資訊’並設定於資訊設定部4。 設定於資訊設定部4之資訊4a，包含例如利用過去 之河川的水災狀況資訊或過去之統計資料來決定之閥値、 -14- (11) 1242145 考慮防洪對象區域之河川的狀況等之訊息、以及危險場所 等之人員配置場所及配備人數等資訊在內。 其次，氣象資訊取得部1從氣象衛星HIMAWARI11 、網際網路1 2、及氣象資訊提供契約機關1 3之至少其中 之一接收到氣象資訊時，會將接收到之氣象資訊提供給支 援處理部5。 另一方面，補助資訊取得部2若取得含有雨量計1 4 檢測到之降雨量資訊、水位計1 5檢測到之水位資訊、以 及其他水災相關之各種資訊等在內之補助資訊時，會將代 表雨量計1 4或水位計1 5等之檢測機器設置場所的設置場 所辨識資料、及接收到之補助資訊提供給補助資訊處理部 3。又，<雨量計1 4及水位計1 5係爲了依據下水道幹線1 6 之水位資訊（或河川之水位資訊）來實施防洪而預先設置於 必要場所。 補助資訊處理部3在從補助資訊取得部2接收到補助 資訊時，會將補助資訊轉換成使用者期望之用以擬定防洪 支援資訊的支援資訊’將支援資訊記錄資訊記錄部8 ’並 提供給支援處理部5。 其次，支援處理部5會執行第5圖所示之處理。 步驟S 1中，支援處理部5之氣象判斷部5 1會判斷從 氣象資訊取得部1接收之氣象資訊是否爲例如大雨警告、 大雨警報、洪水警告、及洪水警報等防洪對象之氣象資訊 〇 判斷爲防洪對象之氣象資訊時，步驟S 2 a〜S 2 d中， -15- (12) 1242145 氣象判斷部5 1會決定氣象資訊之種別爲大雨警告、大雨 警報、洪水警告、及洪水警報之其中那一種。第5圖中， 記錄著判斷是否爲種別資料D 1之步驟S 2 a、及判斷是否 爲種別資料D4之步驟S2d，而省略判斷是否爲種別資料 02、〇3之步驟8213、32(：。 又，亦可在步驟S 1判斷爲大雨警告、大雨警報、洪 水警告、及洪水警報之其中那一種。氣象資訊爲大雨警告 、大雨警報、洪水警告、及洪水警報之其中任一種時，會 從氣象判斷部5 1之處理進入必要性判斷部5 2之處理。 步驟S 3〜S 5中，必要性判斷部5 2會從資訊記錄部8 取得最接近現在時刻之時間帶的例如包含降雨量資訊等在 內之支援資訊，並將支援資訊和設定於資訊設定部.4之閥 値進行比較，判斷是否需要實施防洪活動。 例如’降雨量超過閥値時、或降雨量位於複數閥値所 規定之範圍內時，會判斷必須實施防洪活動。 更具體而Η，若閥値設定爲5 0 m m / h以上時，必要性 判斷部52在降雨量資訊所代表之降雨量超過閥値5〇mnl/h 時，會判斷需要實施防洪活動。 利用複數閥値設定範圍時，必要性判斷部5 2在降雨 量屬於該範圍內時會會判斷需要實施防洪活動。 必要性判斷部5 2若判斷爲需要採取防洪活動時，會 從氣象判斷部5 1之處理進入必要性判斷部5 2之處理。 步驟S 6中’判斷爲必須採取防洪活動時，擬定部5 3 會從資訊設定部4讀取對應該閥値之訊息、人員配置場所 -16- (13) 1242145 '配備人數等資訊，並實施將讀取之資訊寫入預先登錄之 基礎畫面資料的特定區域內等編輯處理，例如，擬定畫面 顯示用防洪支援資訊。 第6圖係擬定之防洪支援資訊的實例。 又’擬定部5 3亦可不使用基礎畫面資料，而只利用 訊息、人員配置場所（X A、X B、X C )、配備人數等之資 訊以特定格式來擬定畫面顯示用防洪支援資訊。自治體預 先決定配備人數時，亦可由擬定部5 3擬定只代表是否有 現場人員配置要求之防洪支援資訊。 又’擬定部5 3亦可在畫面顯示用防洪支援資訊內追 加現在針對必要區域發佈之氣象資訊、用以擬定防洪支援 資訊之降雨量資訊、及閥値等資訊。 擬定防洪支援資訊後，會從擬定部5 3之處理進入輸 出部5 4之處理。 步驟S 7中，輸出部5 4將擬定之防洪支援資訊記錄於 圖上未標示之適當記錄手段後，會實施顯示控制，將顯示 控制之結果記錄於圖上未標示之圖像記憶體等，在例如上 述第6圖所示狀態將防洪支援資訊顯示於顯示裝置6。 步驟S8中’在經過特定時間後，輸出部54會確判斷 是否有確認指示被輸入。 沒有確認指示被輸入時，步驟S9中，輸出部54會輸 出警報’並再度移至步驟S 8執行處理。 有確認指示被輸入時，步驟S ；! 〇中，輸出部54會判 斷已結束處理。未結束處理時，會以特定周期重複執行上 -17- 1242145 (14) 述步驟s 1以下之處理。 上述支援處理部5之處理的執行上’係以接收到如大 雨警告、大雨警報、洪水警告、及洪水警報之防洪對象的 氣象資訊爲條件。 然而，即使末接收到防洪對象之氣象資訊時，防洪對 象區域有時亦會降下大雨。此種狀況時，則需判斷爲需要 實施防洪活動並擬定防洪支援資訊。 爲了在未接收到防洪對象之氣象資訊時能以支援資訊 來判斷是否需要實施防洪活動，即使在上述第5圖所示之 步驟S 1或步驟S2d判斷爲否時，亦會進入步驟S3並執行 處理。其次，例如將降雨量及閥値進行比較等，判斷是否 需要實施防洪活動。 利用此方式，即使未接收到防洪對象之氣象資訊，但 判斷需要實施防洪支援時，會擬定防洪支援資訊並顯示。 又，支援處理部5亦可依據其他防洪處理程式擬定第 7圖所示之畫面顯示用防洪支援資訊。 第7圖所示之畫面顯示用防洪支援資訊實例中，會將 包含發佈時間及發佈之氣象內容在內的氣象資訊2 3讀取 至預先登錄之基礎畫面資料22的特定區域，且以圖形顯 示現在之降雨量及水位。 水位圖中，境界線2 1 a係代表上述第4圖之警戒水位 21。境界線24係代表上述第4圖之累計降雨量的 下限線水位。境界線2 5係代表上述第4圖之累計降雨量 3 5 m m的上限線水位。 -18- 1242145 (15) 本竇施形態具有如下所示之效果。 例如，如第8圖所示，只依1 〇 〇次下雨之以往氣象資 訊來實施防洪活動。第8圖中，1 〇〇次下雨當中，發佈大 雨警報且發生水災之次數爲5次，發佈大雨警報卻未發生 水災之次數爲1 5次。 只利用氣象資訊之以往的防洪活動，會針對20次大 雨警報實施2 0次防洪活動。 因此，2 0次防洪活動中，只有5次防洪活動是必要 的（準確），1 5次之防洪活動爲不必要（不準確）。 另一方面，上述第8圖之狀態下，有下雨但未發佈大 雨警報之次數爲8 0次，該8 0次當中發生1次水災。此時 、，因在未實施防洪活動之狀態下發生I次水災，1次水災 完全被忽略。 第9圖係附加支援資訊來擬定防洪支援資訊並顯示日寺 之實例圖。第9圖係利用上述第5圖之步驟S1〜S5之處 理來判斷是否需要實施防洪活動之結果實例。 第9圖中，利用發佈之大雨警報、及支援資訊判斷是 否需要實施防洪活動時，會判斷發佈大雨警報且發生水災 之5次的全部下雨都需要實施防洪活動。因此，因係依據 大雨警報及支援資訊來判斷防洪活動之必要性並實施防洪 活動，故會針對發佈大雨警報且發生水災之5次下雨皆實 施防洪活動、及配備必要人員，而可避免水災並確保安全 狀態。 如上述第8圖所示，只以大雨警報來判斷防洪活動之 -19- 1242145 (16) 必要性時’發佈大雨警報且未發生水災之防洪活動必要性 之判斷上，不準確的次數爲1 5次。 相對於此，依據大雨警報及支援資訊來判斷防洪活動 之必要性時，發佈大雨警報且未發生水災之1 5次當中， 有5次會判斷成危險而判斷必須實施防洪支援（防洪支援 資訊輸出）’ 1 0次會被判斷成無水災之危險性而未輸出防 洪支援資訊。 結果’發佈大雨警報但未發生水災之1 5次當中，有 1 〇次會判斷成安全，而實際上亦可該判斷結果相同，並 未發生水災。如上所示，利用依據大雨警報及支援資訊來 判斷防洪活動之必要性，可將不準確次數從1 5次減至5 次。、 又’第9圖之結果係將統計演算所得之上述第4圖之 危險水位當做閥値來進行判斷時之實例。變更閥値之設定 方法可以獲得更正確之判斷結果。 如第8圖所示，只利用大雨警報來判斷防洪活動之必 要性時，未發佈大雨警報卻發生水災之未注意到危險的情 形發生1次。 相對於此，如第9圖所示，依據大雨警報及支援資訊 來判斷防洪活動之必要性時，未發佈大雨警報但處於危險 狀態，而判斷成需要採取防洪活動之次數爲1次，故不會 發生未注意到危險之情形。 未發佈大雨警報且未發生水災時，亦即，降雨量較少 時，會和以往相同，針對7 9次判斷成無發生水災之危險 -20- 1242145 (17) ，且未輸出防洪支援資訊。 如以上所示，本實施形態時，除了大雨警告、大雨警 報、洪水警告、及洪水警報等氣象資訊以外，尙會利用補 助資訊及支援資訊當中之至少其中一方來判斷防洪活動之 必要性。利用此方式，可大幅降低防洪活動之不準確性及 被忽略的情形。 又，本實施形態時，因爲會輸出可促成符合防洪活動 之人員配備的防洪支援資訊，故很容易即可預先防止水災 之發生。 又’本實施形態時，若判斷需要採取防洪活動，會顯 示訊息及防洪活動配備人數，可參考適當之訊息針對防洪 活動_h之必要場所配備適當人數。 又’本實施形態時，係利用降雨量資訊及水位資訊等 支援資訊（或補助資訊）來決定閥値，然而，亦可以利用例 如以計算目的之達成度爲目的之評估函數來決定閥値。 M $D ’防洪活動之1次現場出動成本x爲3萬日元， 例如堆土等之1次現場出動的作業成本y爲1 0萬曰元， 發主水災時之1次水災作業成本z爲1 0 0萬日元，即可利 用成本來定義評估函數。 上·述第8圖中，結果「（ a)準確◎」的發生次數爲5 次’成本種類爲「y」。因此，結果「（a)準確◎」之評估 値爲5 0萬曰元。 結果「（b)不準確△」的發生次數爲15次，結果「（b) 不準確△」之成本種類爲「x」。因此，結果「（b)不準確 -21 - 1242145 (18) △」之評估値爲4 5萬日元。 結果「（c)忽略X」的發生次數爲1次，結果「（c)忽 略X」之成本種類爲「z」。因此，結果「（c)忽略X」之 評估値爲1 〇 〇萬日元。 結果「（d)準確〇」的發生次數爲79次，結果「（d)準 確〇」之成本爲〇。結果，結果「（d)準確〇」之評估値爲 0日元。 上述第8圖之評估値的合計爲1 95萬日元。 第8圖之結果「（a)準確◎」、結果「（b)不準確性△ 」、結果f (c)忽略X」、及結果「（d)準確〇」之各發生 次數爲A、B、C、及D，成本x、y、z之値分別爲χ、γ 、Ζ，則可以下式（1 ).來表示評估函數J。 J = A xY + B χΧ + C x Z ··· (1) 例如，有如上述第4圖所示之累計降雨量資訊，將閥 値設定成累計降雨量20mm，並假設防洪活動之執行結果 如第1 0圖所示。 針對第1 〇圖利用上述式（1 )之評估函數來計算成本値 ，可得到如下所示之結果。1242145 (1) 发明 Description of the invention [Technical field to which the invention belongs] The present invention relates to a flood support device and program for supporting flood control activities for the purpose of avoiding floods, and a flood support method. [Prior technology] During the flood prevention activities of the sewers or rivers, the staff shift system is determined according to the warnings or alarms issued by the Meteorological Bureau. The shifts will arrive at the scene according to the conditions of the sewers or rivers. In the past, there have also been proposed disaster prevention support devices that obtain rainfall information and use it for disaster prevention activities. For example, the disaster prevention information providing device of Patent Document 1 includes: a disaster prevention information storage unit for storing disaster prevention information such as magnitude information, seismic intensity information, and distortion information; a browser-equipped machine held by a user; and a link to receive A WWW link that is accessed by a user with a browser-equipped machine and outputs disaster prevention information reading requests. Next, the disaster prevention information providing device of Patent Document 1 receives a read request from the WWW link unit, determines whether to provide the disaster prevention information of the disaster prevention information memory unit, and when it is determined to be available, provides the disaster prevention information to the user via the WWW link unit. . For example, the disaster prevention information collection and processing system of Patent Document 2 uses the location information of the disaster site as a keyword, and obtains the image information of the registered predicted disaster occurrence site. Second, the disaster prevention information collection and processing system of Patent Document 2 combines and edits the image information and rainfall information, and then stores the edited disaster prevention information in a recording device and publishes it through the information providing department-5- 1242145 (2) It is widely available to users on Internet servers. [Patent Document 1] Japanese Unexamined Patent Publication No. 1 1-1 1 0454 [Patent Document 2] Japanese Unexamined Patent Publication No. 2001-337997 [Summary of the Invention] The meteorological bureau's warning or alarm may not be targeted at the flood prevention area 5 ° Therefore, when the crew's shifts are determined based on the warnings or alarms of the Meteorological Bureau, the shifts that have already been decided may only need to remain on standby without taking any specific flood prevention activities. The above-mentioned traditional disaster prevention information provision device or disaster prevention information collection and processing system will widely provide disaster prevention information prepared based on rainfall information and other information to users' and use it in disaster prevention activities of regional residents. However, traditional disaster prevention information provision devices or disaster prevention information collection and processing systems do not have the function to support the prevention of general flood prevention activities such as floods caused by watercourses or rivers. Generally speaking, flood prevention activities include publicity activities for regional residents, Various activities such as surveillance, site security such as bulldozing, and relocation of residents from dangerous places to safe places. However, because traditional disaster prevention information provision devices or disaster prevention information collection and processing systems only provide users with disaster prevention information, measures to avoid floods and other flood prevention activities cannot be effectively implemented. Therefore, traditional flood prevention activities for floods have traditionally been It refers to the heavy rain warning, heavy rain warning, flood warning, and flood warning issued by specific organizations such as Meteorology-6- (3) 1242145 for the region as indicators. Secondly, the traditional general flood prevention activities are based on the information represented by any of the heavy rain warning, heavy rain warning, flood warning, and flood warning, and past experience, to estimate the place where the flood occurred, and then deploy the necessary personnel to the flood prevention. Necessary places. However, when only using the meteorological information as an indicator for flood prevention activities, because the release of meteorological information is targeted at a wide range of areas, if there is no heavy rain in the place where the actual flood prevention activities are taking place, it will occur Flood prevention activities are carried out in places such as flood prevention activities, and there is a problem of implementing flood prevention activities that waste resources. In view of the above-mentioned circumstances, the object of the present invention is to provide a flood support device and program, and a flood support method, which can prevent flood prevention activities in places where flood prevention activities are not needed, and can implement flood prevention activities in places where flood prevention activities are needed. Effective flood prevention activities. The following is a description of specific means for implementing the present invention. The first aspect of the present invention includes: means for obtaining meteorological information; necessary support information means for obtaining flood control support; and judging the necessity of flood control activities based on meteorological information and support information, and when planning for flood prevention activities is required Necessary personnel are equipped with information to support and deal with means of flood control support δΤΙ; flood control support device. The second aspect of the present invention includes: means for obtaining meteorological information; means for obtaining subsidy information related to floods; means for processing necessary subsidy information for obtaining flood control support based on the aid information; and means for meteorological information and (4) 1242145 turtle support information To determine the necessity of flood control activities and to determine the need for flood control activities, draft a flood control including the necessary staffing information for flood control activities: support processing methods to support information; and flood control support devices. 'The third aspect of the present invention includes: means for obtaining analysis target information including information for judging the past and past of the flood prevention and the degree of danger corresponding to the past information; means for obtaining the current point information for judging the flood prevention; And according to the forecasting means obtained by implementing data mining for the analysis target information, and the current point information, the means of drafting flood control support information that represents the degree of danger at the time point; the flood control support device. The fourth aspect of the present invention includes: means for obtaining teacher information including past information for judging flood prevention and the degree of danger corresponding to the past information; means for obtaining current point information for judging flood prevention; Teachers implement neural networks such as learning and current point information, and develop means for flood control support information that represent the degree of danger at the time point; a flood control support device. [Embodiment] Below φ, an embodiment of the present invention will be described with reference to the drawings. -"In the following description, the same elements will be given the same symbols, and their explanations will be omitted. * (First Embodiment) _ 1 is a block diagram of an example of a flood control support device according to this embodiment. The flood control support device 101 has a weather information acquisition unit 1, subsidy information-8- 1242145 (5) ear 5 acquisition unit 2, subsidy information processing unit 3, information setting unit 5, display device 6, and input devices 7a, 7b . The meteorological information acquisition unit 1 is used to acquire meteorological information, such as information indicating whether a heavy rain warning, a heavy rain warning, and a flood warning will occur. For example, the weather information acquisition unit 1 uses HIMAWARI1 1 to receive weather information. In addition, for example, the ministry 1 may receive weather information from the government's weather information provider through the Internet. In addition, for example, weather information acquisition department 1 center and contract agency provide contract agency information. The subsidy information acquisition unit 2 is used to obtain subsidy information related to meteorological floods. For example, the subsidy information includes rainfall data detected by a rain gauge 14 installed in the area targeted for flooding. For example, the subsidy information includes water level information such as the use of water level gauges for sewer lines and rivers. In addition, subsidy information may be included in connection with floods. The subsidy information can be any of rainfall information, sewer trunks, and river level information, including two or more types. The subsidy information processing department 3 is used to implement the necessary support to convert the subsidy information obtained by using supplementary dollars into flood control activities. 4. Support processing. Meteorological information is a flood warning. The purpose is to obtain weather information from meteorological satellites. The Internet 1 2 can receive gas information directly from the meteorological business 3. The river and other defense information (rainfall information) 1 5 Various other information detected One of the online water level information can also assist in the processing of support information required by the information acquisition department-9-1214545 (6). The information setting section 4 sets necessary information for drawing up flood prevention support information. The support processing unit 5 will formulate necessary flood prevention support information for the flood control activities. The display device 6 will display the flood protection support information formulated by the support processing unit 5. The input devices 7 a and 7 b are necessary information for, for example, the purpose of formulating the flood protection support information. And input of various instructions. Fig. 2 is a relationship diagram of the locations of the sewer lines 16 (or rivers) of the flood control target area AB 1 in the county A and the B area, and the locations of the rain gauge 14 and water level gauge 15 of the subsidy information detection system. In the example in FIG. 2, two rain gauges 14 are installed at necessary places of the flood control target area AB !. In addition, two water level gauges 15 are installed along the water main line 16 flowing below the flood protection target area A B 1. The meteorological information acquisition unit 1 receives meteorological information such as a heavy rain warning, a heavy rain alarm, a flood warning, and a flood alarm from the meteorological information contracting agency 1 3, and will provide the received meteorological information to the support processing department 5 ° subsidy information The acquisition unit 2 will receive, for example, the use of a rain gauge! 4 Supported information including detected rainfall information, water level gauge 1 5 detected water level information, and other flood related information. In order to use the support information to formulate support information, the support information and detection The identification information of the installation location of the subsidy information is provided to the subsidy information processing section 3. -10- (7) 1242145 The subsidy information processing unit 3, for example, when the subsidy information includes water level information, will identify the data for each installation place, and convert the subsidy information into representative water level changes at specific times (for example, every 5 minutes). The water level condition information, and the support information including the water level information and the water level condition information for each specific time obtained according to the time series are recorded in the information recording section 8. When the water level information is processed by the method shown above, the setting location identification data is used as a keyword, and support information including water level information, water level condition information, and other information is recorded in the information recording section 8. When the subsidy information processing unit 3 includes, for example, rainfall information, the subsidy information processing unit 3 converts the subsidy information into accumulated rainfall information that starts to accumulate at the beginning of rain or the maximum value for each specific time when the subsidy information includes the rainfall identification information. Rainfall information, and support information including the accumulated rainfall information obtained or the maximum rainfall information for each specific time will be recorded in the information recording section 8 according to the time series. When processing the rainfall information as shown above, the setting identification data is used as a keyword, and the supporting information including rainfall information, accumulated rainfall information, maximum rainfall information, and other information is recorded in the information. Recording section 8. The information setting section 4 sets necessary information 4 a for formulating the flood support information as shown in FIG. 3. In other words, the information setting unit 4 sets the identification information "〇!", "〇2", ... for each setting place, and sets the weather information type data (D1: heavy rain warning, D2: heavy rain alarm, D3: flood warning, D4: flood Alarms), valves, messages, staffing locations, and the number of staff at each facility. In addition, the valve can be set for any one of the rainfall information [mm / h], accumulated rainfall information [111111], maximum rainfall information [mm / h], water level information, water level condition information, and other information. , However, it should be set for information that is most susceptible to floods to sewers or rivers. For example, the rainfall can be divided into "50 mm / h ~ 9 9 mm / h", "100 mm / h ~ 1 49 mm / h", "more than 150 mm / h", and One type of information is setting plural valves 値. For example, the valve 値 may be set for two or more kinds of information such as accumulated rainfall information and water level information. The valve can also be determined based on the rainfall information in the past when the flood occurred, the water level information of the main waterway in the past when the flood occurred, and the water level information of the river in the past when the 7_K disaster occurred. For example, in the past when a flood occurred at a rainfall of 50 mm / h, it is also possible to set 5 Omm / h as the valve of the rainfall. However, since it is necessary to increase the number of employees in response to an increase in rainfall, it is best to set the valve 针对 for plural rainfalls, and set the number of employees for each valve 値. In addition, when floods have not occurred in the past, the valve 値 can also be set based on the planned rainfall calculated during the design of the facility. The information set in the information setting section 4 is different because each sewer main line or river and each installation place identification data "0 1", "02", ... are different. It is possible to describe the correspondence to each sewer main line or river, and installation place identification data. content. Because there are often more than one dangerous places around the sewer line or river -12- (9) 1242145 Dangerous places ’, it is possible to effectively set multiple staffing places for each sewer line or river. The number of staff is set for each staffing location. In some cases, local governments may only implement flood control at specific locations on one river. At this time, the setting place identification data may be omitted in the information 4a. If flood prevention activities are implemented regardless of any kind of meteorological information, the kind of meteorological information may be omitted in Information 4a. That is, the information setting unit 4 only needs to set any one of the staffing place and the number of employees in each staffing place with respect to any one of the setting place identification data, valves, and messages. In addition, for example, when setting the valve 资讯 based on rainfall information, water level information of a sewer line, and river level information, the valve 値 may also be set based on calculations based on past statistical data. Figure 4 is an example of past statistics. It is a relationship diagram of the past rainfall information [m m / h] or the past accumulated rainfall information [m m], the sewer main line past water level information, or the river past water level information recorded on the support processing unit 5. In addition, it is possible to determine the valve valve in consideration of the past flood situation shown in FIG. 4. In Figure 4, when the accumulated rainfall reaches 2 9 [m m], the upper limit of the approximate curve using the past statistical data is to reach the warning water level 21 of the sewer line 0.6. Therefore, for example, the valve 値 is set to a cumulative rainfall of 29mm and a water level of 0.6m, and the staffing is performed when the accumulated rainfall or water level exceeds the valve 値. The support processing unit 5 can also perform logical calculations and judgments by itself, and output flood prevention support information. When the support processing unit is constituted by, for example, a CPU, the flood processing program can also be used to perform specific processing and formulate flood support information. The support processing unit 5 has a functional weather judgment unit 5 1. A necessity judgment unit 5 2, a planning unit 5 3, and an output unit 54. The meteorological judgment unit 51 judges whether the meteorological information received by the meteorological information acquisition unit 1 is meteorological information targeted for flood prevention. Meteorological information such as a heavy rain warning, a heavy rain alarm, a flood warning, and a flood alarm is defined as the meteorological information of the flood prevention object. The necessity determination unit 5 2 refers to the support information such as rainfall information, which is closest to the current time when the information is recorded by the information recording unit 8 when using the weather information determined by the weather determination unit 51 as a flood prevention object. When the valve of the information setting unit 4 is judged, it is judged that flood prevention activities must be implemented. The planning unit 5 3 When the necessity judgment unit 5 2 judges that flood control activities are necessary, it will read the information 'staffing place, and number of personnel data' from the information setting unit 4 and implement editing processing in the necessary format to formulate the flood prevention Support information. For example, 'flood control support information' can also be used for screen display. The output unit 54 outputs the flood prevention support information prepared by the planning unit 53 to the display device 6. First, the input devices 7a'7b input necessary information for formulating flood control support information 'and set it in the information setting section 4. The information 4a set in the information setting unit 4 includes, for example, a valve determined based on past flood situation information or past statistical data, -14- (11) 1242145 considering the status of rivers in areas targeted for flood control, etc., It also includes information on staffing locations and number of personnel in hazardous locations. Next, the meteorological information acquisition unit 1 provides meteorological information to the support processing unit 5 when it receives meteorological information from at least one of the meteorological satellite HIMAWARI11, the Internet 12, and the meteorological information providing contract agency 13. . On the other hand, if the subsidy information acquisition unit 2 obtains subsidy information including rainfall information detected by the rain gauge 1 4, water level information detected by the water level gauge 15, and various other information related to floods, it will The installation location identification data representing the installation location of the testing equipment such as the rain gauge 14 or the water level gauge 15 and the received subsidy information are provided to the subsidy information processing section 3. In addition, < Rain gauge 14 and Water level gauge 15 are installed at necessary places in advance to implement flood prevention based on water level information (or river level information) of the sewer main line 16. When receiving the subsidy information from the subsidy information acquisition unit 2, the subsidy information processing unit 3 converts the subsidy information into the support information that the user expects to be used to formulate the flood prevention support information. Support processing department 5. Next, the support processing unit 5 executes the processing shown in FIG. 5. In step S1, the meteorological judgment unit 51 of the support processing unit 5 judges whether the meteorological information received from the meteorological information acquisition unit 1 is meteorological information of a flood prevention object such as a heavy rain warning, a heavy rain alarm, a flood warning, and a flood alarm. When the weather information is the target of flood control, in steps S 2 a to S 2 d, -15- (12) 1242145 The weather judgment unit 51 will determine the type of weather information as heavy rain warning, heavy rain warning, flood warning, and flood warning. Which one. In FIG. 5, steps S 2 a for judging whether it is the category data D 1 and step S 2 d for judging whether it is the category data D4 are recorded, and steps 8213 and 32 (:) that judge whether it is the category data 02 and 03 are omitted. In step S1, one of the heavy rain warning, heavy rain warning, flood warning, and flood warning may be determined. When the weather information is any of the heavy rain warning, heavy rain warning, flood warning, and flood warning, The process of the weather judgment unit 51 enters the process of the necessity judgment unit 52. In steps S3 to S5, the necessity judgment unit 52 obtains the time zone closest to the current time from the information recording unit 8 including, for example, rainfall. Support information including information, and compare the support information with the valve set in the information setting section.4 to determine whether flood control activities are needed. For example, when the rainfall exceeds the valve level, or the rainfall is located in multiple valve positions. If it is within the specified range, it will be judged that it is necessary to implement flood prevention activities. More specifically, if the valve 値 is set to 50 mm / h or more, the necessity judgment section 52 indicates the rainfall represented by the rainfall information. When it exceeds valve 〇50mnl / h, it is judged that flood prevention activities are necessary. When setting a range using a plurality of valve 必要, the necessity judgment unit 52 determines that flood prevention activities are necessary when rainfall falls within this range. Judgment of necessity If the unit 52 determines that it is necessary to take flood control activities, it will proceed from the process of the weather judgment unit 51 to the process of the necessity determination unit 52. When it is determined in step S6 that flood control activities must be taken, the planning unit 5 3 will start from The information setting section 4 reads the information corresponding to the valve valve and the staffing location-16- (13) 1242145 'Equipment number and other information, and edits the read information into a specific area of the pre-registered basic screen data. Processing, for example, drafting of flood control support information for screen display. Figure 6 is an example of the planned flood control support information. Also, the drafting section 5 3 may not use the basic screen data, but only use information and staffing places (XA, XB, XC), the number of staff, etc. are prepared in a specific format for the screen display flood control support information. When the local government determines the number of staff in advance, the planning department 5 3 The table indicates whether there is flood control support information required by the on-site staffing requirements. The planning section 53 can also add to the screen display flood control support information meteorological information now released for necessary areas, rainfall information used to formulate flood control support information, and Valves and other information. After drafting the flood control support information, it will go from the processing of the planning section 53 to the processing of the output section 54. In step S7, the output section 5 4 records the planned flood control support information on the map that is not marked as appropriate. After the recording means, display control will be implemented, and the results of the display control will be recorded in the unmarked image memory, etc., and the flood prevention support information will be displayed on the display device 6 in the state shown in Figure 6 above. Step S8 ' After a certain time has elapsed, the output unit 54 determines whether or not a confirmation instruction has been input. When no confirmation instruction is input, the output unit 54 outputs an alarm 'in step S9 and moves to step S8 to execute the process again. When a confirmation instruction is input, in step S;! 〇, the output unit 54 judges that the processing has ended. When the processing is not completed, the above-mentioned steps s 1 to -17-1242145 (14) are repeatedly executed at a specific cycle. The execution of the processing by the above-mentioned support processing unit 5 is conditional upon receiving weather information of a flood prevention object such as a heavy rain warning, a heavy rain warning, a flood warning, and a flood warning. However, even when the weather information of the flood prevention object is not received, the area of the flood prevention object sometimes drops heavy rain. In such a situation, it is necessary to judge the need for flood prevention activities and develop flood prevention support information. In order to judge whether the flood prevention activities need to be implemented with the support information when the weather information of the flood prevention object is not received, even if the judgment in step S 1 or step S2d shown in the above figure 5 is no, it will proceed to step S3 and execute deal with. Secondly, for example, comparing rainfall and valve 値 to determine whether flood prevention activities are needed. With this method, even if the weather information of the flood prevention object is not received, but it is determined that flood prevention support is needed, the flood prevention support information will be drawn up and displayed. In addition, the support processing unit 5 can also prepare flood control support information for screen display shown in FIG. 7 based on other flood control processing programs. In the example of the flood prevention support information for screen display shown in Fig. 7, the weather information including the release time and the released weather content is read to a specific area of the pre-registered basic screen data 22 and displayed graphically. Current rainfall and water levels. In the water level diagram, the boundary line 2 a represents the warning water level 21 in the above-mentioned FIG. 4. The boundary line 24 is the water level of the lower limit line of the accumulated rainfall shown in FIG. 4 above. The boundary line 2 5 represents the upper limit line water level of the accumulated rainfall of 3 5 m in Figure 4 above. -18- 1242145 (15) This sinus application has the following effects. For example, as shown in Figure 8, flood prevention activities were implemented based only on past weather information for 1,000 rains. In Figure 8, among the 1,000 rains, heavy rain alarms were issued and floods occurred five times, and heavy rain alarms were issued without floods 15 times. For past flood control activities using only weather information, 20 flood control activities were implemented for 20 heavy rain alarms. Therefore, out of 20 flood control activities, only 5 flood control activities are necessary (accurate), and 15 flood control activities are unnecessary (inaccurate). On the other hand, in the state shown in Figure 8 above, the number of times when there was rain but no heavy rain alarm was issued was 80 times, and one of the 80 times occurred a flood. At this time, because of the occurrence of one flood in the state where no flood prevention activities were implemented, one flood was completely ignored. Fig. 9 is an example diagram of additional support information to formulate flood prevention support information and display the temple. Fig. 9 is an example of the result of using steps S1 to S5 in Fig. 5 to determine whether flood control activities are needed. In Figure 9, using the issued heavy rain alarm and supporting information to determine whether flood prevention activities are needed, it will be determined that all the five rains that occurred when the heavy rain alarm was issued and floods require flood prevention activities. Therefore, since the necessity of flood prevention activities is determined and implemented based on the heavy rain warning and support information, flood prevention activities will be implemented for the five rains that occur when the heavy rain warning is issued and the necessary personnel will be provided to prevent floods. And ensure a safe state. As shown in Figure 8 above, only the heavy rain alarm is used to judge the flood prevention activities. -19- 1242145 (16) When necessary, the number of inaccurate judgments when a heavy rain alarm is issued without the need for flood prevention activities is 1 5 times. On the other hand, when judging the necessity of flood control activities based on heavy rain alarms and support information, 5 of the 15 times when heavy rain alarms are issued without flooding are judged as dangerous and it is judged that flood prevention support must be implemented. ) '10 times will be judged as no danger of floods and no flood prevention support information is output. As a result, of the 15 times when a heavy rain alarm was issued but no flood occurred, 10 times were judged to be safe, but in fact, the judgment result could be the same and no flood occurred. As shown above, the use of heavy rain alarms and support information to judge the necessity of flood control activities can reduce the number of inaccurate times from 15 to 5. The result of Fig. 9 is an example when the dangerous water level of the above Fig. 4 obtained as a result of statistical calculation is used as a valve to judge. Changing the setting method of the valve can obtain more accurate judgment results. As shown in Figure 8, when only the heavy rain alarm is used to judge the necessity of flood control activities, the unnoticed danger of flooding occurs without issuing a heavy rain alarm once. In contrast, as shown in Figure 9, when the necessity of flood prevention activities is judged based on the heavy rain alarm and support information, the heavy rain alarm is not issued but is in a dangerous state, and the number of times that it is necessary to take flood prevention activities is determined, so it is not necessary. Situations where danger is not noticed can occur. When no heavy rain alarm is issued and there is no flood, that is, when there is less rainfall, it will be the same as in the past, and it is determined that there is no danger of flooding for 7 or 9 times -20-1242145 (17), and no flood prevention support information is output. As shown above, in this embodiment, in addition to the meteorological information such as heavy rain warning, heavy rain alarm, flood warning, and flood warning, at least one of the supplementary information and support information is used to determine the necessity of flood prevention activities. Using this method can greatly reduce the inaccuracies and neglect of flood prevention activities. Furthermore, in this embodiment, flood prevention support information is provided which can be provided by personnel who can contribute to flood prevention activities, so it is easy to prevent floods in advance. Also, in this embodiment, if it is determined that flood control activities are needed, a message and the number of persons equipped with the flood control activities will be displayed. You can refer to the appropriate information for the necessary places for the flood control activities_h. In this embodiment, the valve 値 is determined using supporting information (or subsidy information) such as rainfall information and water level information. However, the valve 値 can also be determined using an evaluation function for the purpose of calculating the achievement degree of the purpose. M $ D 'The cost of one on-site dispatch of flood prevention activities x is 30,000 yen. For example, the cost of one on-site dispatch of a pile of soil is 100,000 yen, and the cost of one flood operation when a flood occurs is z For one million yen, the cost can be used to define the evaluation function. As shown in Fig. 8 above, the number of occurrences of "(a) accurate ◎" is 5 times and the cost type is "y". Therefore, the evaluation of the result "(a) accurate ◎" is 500,000 yuan. As a result, the number of occurrences of "(b) inaccurate △" was 15 times, and the result "(b) inaccurate △" was "x". Therefore, the evaluation of the result "(b) Inaccurate -21-1242145 (18) △" is 40,000 yen. As a result, "(c) Ignore X" occurred once, and as a result, "(c) Ignore X" cost type was "z". Therefore, the evaluation of the result "(c) Ignore X" is 10 million yen. As a result, the number of occurrences of "(d) accurate 0" was 79, and the result "(d) accurate 0" had a cost of 0. As a result, the evaluation of the result "(d) Accuracy 0" was 0 yen. The total of the evaluation cards in Figure 8 above is 19.5 million yen. The number of occurrences of the results "(a) accurate ◎", result "(b) inaccuracy △", result f (c) ignore X ", and result" (d) accurate 0 "in Fig. 8 are A, B , C, and D, and the cost of x, y, and z are respectively χ, γ, and Z. Then, the evaluation function J can be expressed by the following formula (1). J = A xY + B χχ + C x Z (1) For example, if there is accumulated rainfall information as shown in Figure 4 above, set valve 値 to accumulate rainfall of 20mm, and assume that the implementation results of flood control activities are as follows Figure 10 shows. For the figure 10, using the evaluation function of the above formula (1) to calculate the cost 値, the following results can be obtained.
J20 = A xY + Β χΧ + C xZ =6 xl0 + 35 x3 + 0 x10 0 =1 65(萬日元） 將上述式（1 )之評估函數計算所得之評估値當做支援 資訊，針對該支援資訊設定閥値’並依據該設定之閥値擬 定畫面顯示用防洪支援資訊並輸出’可實現比以往更安全 -22- 1242145 (19) 且更高評估値（低成本）之防洪活動。 第1 〇圖中，不準確之次數雖然較多，但成本較便宜 〇 又，評估函數並非只可應用於支援資訊之取得，亦可 應用於求取其他資訊。例如，如上述第4圖所示，利用累 計降雨量計算水位等，將利用統計資料以評估函數計算所 得之値當做閥値使用亦可。 (第2實施形態） 本賃施形恶係針提供以下彖雨時之下水道或河川之 防洪爲目的之有效資訊來支援防洪活動之防洪支援裝置進 :行說明。， 第1 1圖係本實施形態之防洪支援裝置之構成實例的 方塊圖。 防洪支援裝置1 〇 2係依據防洪對象區域內之過去降雨 量資訊、過去河川或下水道幹線內之水位資訊、以及氣象 局之氣象資訊（數値資料）實施資料挖掘，並擬定以數値轰 示針對任何監視點之人員配置的危險度、或緊急處理度之 防洪支援資訊。 利用該防洪支援資訊，可依據防洪對象區域之降雨量 或水位等觀測狀況之把握、及過去經驗來採取措施，而可 實現高精度且迅速之人員配置活動。 又’本實施形態時，資料控掘之手法係針對利用類神 經網路之圖形辨識時進行說明，然而，亦可以採用例如多 -23- 1242145 (20) 變量解析及各種統計手法等。 防洪支援裝置1 〇 2具有資料挖掘部；3 〇、教師資料取 得部3 1、氣象資訊取得部3 2、補助資訊取得部3 3、補助 資訊處理部3 4、以及輸出部3 5。 資料挖掘部3 0具有利用類神經網路之圖形辨識的機 月b 。 教師資料取得部3 1會從過去資訊記錄部3 6取得資料 挖掘部3 0用以執行類神經網路之學習的教師資料，並提 供給資料挖掘部3 0。 氣象資訊取得部3 2會從例如氣象局接收大雨警報等 氣象資訊，並提供給資料控掘部3 0。 補助:資訊取得部3 3會取得例如雨量計丨4檢測到之降 雨量資訊、及水位計1 5檢測到之水位資訊等補助資訊， 並提供給補助資訊處理部3 4。 補助資訊處理部3 4會將從補助資訊取得部3 3接收到 之補助資訊轉換成支援資訊，並將支援資訊提供給資料挖 掘部3 0。例如，補助資訊處理部3 4會從降雨量資訊取得 單位時間之降雨量資訊，又，從水位資訊取得單位時間之 水位資訊，並將含有單位時間之降雨量資訊及水位資訊雙 方在內之支援資訊提供給資料控掘部3 0。 資料控掘部3 0在利用教師資料實施學習後，輸入含 以氣象資訊取得部3 2取得之氣象資訊、及以補助資訊處 理部3 4取得之支援資訊在內之輸入資訊，並輸出對應輸 入資訊之輸出資訊。 -24- 1242145 (21) 輸出部3 5會依據輸出資訊輸出資料控掘結果。 又，補助資訊取得部3 3亦可不經由補助資訊處理部 3 4而將取得之補助資訊直接提供給資料控掘部3 〇。此時 ’亦可省略補助資訊處理部3 4。又，此時之資料挖掘部 30會輸入含氣象資訊及補助資訊在內之輸入資訊。 又，補助資訊處理部3 4亦可擬定含從補助資訊取得 部3 3接收之補助資訊在內之支援資訊，並將擬定之支援 資訊提供給資料挖掘部3 0。 又’亦可將以補助資訊取得部3 3取得之補助資訊、 及以補助資訊處理部3 4取得之支援資訊雙方提供給資料 挖掘部3 0。此時，資料挖掘部3 0會輸入含氣象資訊、支 援資訊、及補助資訊在內之支援資訊。 第1 2圖係本實施形態之防洪支援裝置丨〇2之資料挖 掘部3 〇的具體實例方塊圖。 式（1 )〜（4)係類神經網路、以及輸入資訊及輸出資訊 之關係圖。 [數式1 ]J20 = A xY + Β χχ + C xZ = 6 xl0 + 35 x3 + 0 x10 0 = 65 (ten thousand yen) Use the evaluation calculated by the evaluation function of the above formula (1) as support information. Set valve 値 ', and based on the set valve 値, prepare flood display support information for screen display and output', which enables safer -22-1242145 (19) and higher evaluation 値 (low cost) flood control activities than before. In Figure 10, although there are many inaccurate times, the cost is cheaper. Also, the evaluation function can be used not only to obtain support information, but also to obtain other information. For example, as shown in Figure 4 above, using accumulated rainfall to calculate water levels, etc., it is also possible to use the data obtained from the calculation of the evaluation function using statistical data as a valve. (Second Embodiment) The following description is provided on the flood control support device for supporting flood control activities by providing effective information for the purpose of flood control of sewers or rivers at the time of rain. Fig. 11 is a block diagram of a configuration example of the flood control support device of this embodiment. The flood control support device 10 is based on the past rainfall information in the flood prevention area, the past water level information in the river or the sewer main line, and the meteorological information (data data) of the Meteorological Bureau to implement data mining. Information on the degree of danger, or emergency response, provided to personnel at any surveillance point. Using this flood prevention support information, measures can be taken based on the grasp of the observation status of rainfall or water level in the flood prevention area and past experience, and high-precision and rapid staffing activities can be achieved. In the present embodiment, the method of data control and excavation is described in the case of pattern recognition using a neural network. However, for example, multi--23-1242145 (20) variable analysis and various statistical methods may be used. The flood control support device 102 has a data mining section; 30, a teacher data acquisition section 3, a weather information acquisition section 3, a subsidy information acquisition section 3, a subsidy information processing section 34, and an output section 35. The data mining unit 30 has a machine identification b using a neural network-like pattern recognition. The teacher data acquisition unit 31 will obtain data from the past information recording unit 36. The mining unit 30 will use the teacher data for performing neural network-like learning and provide it to the data mining unit 30. The meteorological information acquisition unit 32 receives meteorological information such as a heavy rain alarm from the Meteorological Bureau and provides it to the data control and excavation unit 30. Subsidies: The information acquisition unit 33 will obtain subsidy information such as the rainfall information detected by the rain gauge 丨 4 and the water level information detected by the water level gauge 15 and provide it to the subsidy information processing unit 34. The subsidy information processing section 34 converts the subsidy information received from the subsidy information acquisition section 33 into support information, and provides the support information to the data mining section 30. For example, the subsidy information processing unit 34 will obtain the rainfall information per unit time from the rainfall information, and obtain the water level information per unit time from the water level information, and will support both the rainfall information and water level information per unit time. Information provided to the Data Control and Mining Department 30. The data control and excavation unit 30 inputs the input information including the meteorological information acquired by the meteorological information acquisition unit 32 and the support information acquired by the subsidy information processing unit 34 after carrying out learning using teacher data, and outputs corresponding inputs. Information output information. -24- 1242145 (21) The output section 3 5 will output the data control results based on the output information. Further, the subsidy information acquisition unit 33 may directly provide the acquired subsidy information to the data control and excavation unit 30 without passing through the subsidy information processing unit 34. In this case, 'the auxiliary information processing section 34 may be omitted. In this case, the data mining unit 30 inputs input information including weather information and subsidy information. In addition, the subsidy information processing unit 34 may prepare support information including the subsidy information received from the subsidy information acquisition unit 33 and provide the proposed support information to the data mining unit 30. It is also possible to provide both the subsidy information obtained by the subsidy information acquisition unit 33 and the support information obtained by the subsidy information processing unit 34 to the data mining unit 30. At this time, the data mining department 30 will input support information including meteorological information, support information, and subsidy information. Fig. 12 is a block diagram of a specific example of the data excavation unit 30 of the flood control support device 〇2 of this embodiment. Equations (1) to (4) are graphs of neural networks and input and output information. [Equation 1]
Yi (m)=Yi (m) =
(1) …<2) (3) E = 1/2 (yi(3)-di2} -25- 1242145 (22) 式（1 )〜（4)中，XI係代表輸入資訊。m代表類神經網 路3 0a之相數。Y/W代表m相第丨個神經元之活性度。 N m代表m相之神經元數。D;代表各監視點（危險幹線點） 之危險値。 例如’教師貧料係利用：含有過去防洪對象區域內之 降雨量資訊、及下水道幹線內之水位資訊、以及來自氣象 局之氣象資訊（數値資料）在內之輸入資訊（X/CM);以及含 得到該輸入資訊之時點之各監視點危險度（或緊急處理度） 在內之輸出資訊（七）。 更具體而言，教師資料之輸入資訊含有以過去防洪對 象區域內之單位時間的降雨量資訊、過去防洪對象區域內 之‘單·位時間的水位資訊、以、及以「1」代表氣象局發佈大 雨警報時、而以「0」代表未發佈時之資訊在內。 過去之單位時間降雨量爲30 [麵/h]、水位爲5.0 [m]， 且氣象局發佈大雨警報時，教師資料之輸入資訊爲 (30,5.0，1) 〇 又，教師資料之輸出資訊則係利用下述資訊，在獲得 該過去之輸入資訊的時點，針對各監視點，以「1」代表 達到危險水位、或發生水溢出等危險且以「〇」代表未發 生危險之資訊。 在獲得過去之輸入資訊的時點，若只有監視點P2發 生危險時，則教師資料之輸出資訊爲（0，1，...）。 在輸入教師資料時，資料控掘部3 0會利用教師資料 以類神經網路3 0 a實施學習，會重複進行學習，直到輸出 -26- !242145 (23) 資訊及實際神經元輸出之平方誤差小於特定等級爲止。 資料挖掘部3 0會保存由學習所獲得之類神經網路 3 〇 a的構成及加權資料。 資料挖掘部3 0在學習後，若從氣象資訊取得部3 2及 _助資訊處理部3 4接收到現時點之輸入資訊，則會將輸 入資訊輸入至類神經網路3 Oa，並輸出代表現時點之各監 視點危險度的輸出資訊。 類神經網路3 0 a之輸出層會輸出活性等級〇〜1之類 比資訊。 資料控掘部3 0會將該類比資訊當做各監視點之危險 度等級（比率）。又，危險度亦可以百分率來表示。 .、；輸出，部3 5會輸出用以將危險度顯示於畫面上之畫面 顯示用防洪支援資訊。操作員依據依防洪支援資訊顯示於 顯示裝置之各監視點危險度，判斷是否出動至現場。 第1 3圖係利用本實施形態之防洪支援裝置丨〇 2執行 之處理的實例之流程圖。 步驟T 1中，教師資料取得部3 1會參照過去資訊記錄 部3 6，取得教師資料，並提供給資料挖掘部3 〇。 步驟T2中，資料控掘部3 0會依據教師資料執行類神 經網路3 0 a之學習。 步驟T3a中，氣象資訊取得部32會從氣象局取得氣 象資訊’並實施取得之氣象資訊的數値化。 步驟T3 b中，補助資訊取得部3 3會從雨量計及水位 計當中之至少其中之一取得補助資訊。 -27- 1242145 (24) 步驟T 3 c中，補助資訊處理部3 4會將補助資訊轉換 成支援資訊。 又，步驟T3a及步驟T3b、T3c之執行順序可自由設 定。亦即，步驟T3a可在步驟T3b、T3c之前執行，亦可 在之後才執行，亦可在執行步驟T3b、T3c中執行。 步驟T4中，資料挖掘部3 〇會將含有數値化之氣象資 訊及支援資訊在內之輸入資訊輸入至類神經網路3 0 a。 步驟T5中，資料挖掘部3 0會取得對應輸入資訊之輸 出資訊，並提供給輸出部3 5。 步驟T6中，輸出部3 5會執行以將輸出資訊輸出至顯 示裝置爲目的之控制。 本實施形態時，可依據防洪支援對象區域之降雨量資 訊及水位資訊等資訊，提供依據過去經驗之防洪支援資訊 。操作員可參考防洪支援資訊，實施高精度且迅速之人員 配置活動。又，可大幅減少未發生危險之現場巡視等不準 確之防洪活動的次數。 又，本實施形態中，例如，亦可將過去防洪對象區域 之降雨量資訊、過去下水道幹線內之水位資訊、及過去之 防洪活動內容（資料）當做教師資料之輸入資訊，且將各監 視點之過去危險度當做教師資料之輸出資訊。此時，將現 時點之防洪對象區域的降雨量資訊、現時點之下水道幹線 內之水位資訊、及現時點之防洪活動內容當做輸入資訊， 利用類神經網路3 Oa即可得到現時點之輸出資訊。 上述各實施形態之防洪支援裝置101、102實現之機 -28- 1242145 (25) l ’亦可利用程式來實現。 如第1 4圖所示，利用程式3 7實現防洪支援裝置1 〇 i 、1 02之機能時，該程式3 7係記錄於記錄媒體3 8，並被 頃取至電腦3 9。其次，電腦3 9具有之例如C P U等控制單 元會依據程式3 7之內容來執行動作。 又’上述各實施形態之防洪支援裝置1 0 1、1 02中， 只要各構成要素能實現相同之動作，可變更配置，又，亦 可自由組合各構成要素，又，亦可自由分割各構成要素， 或者’刪除數種構成要素。亦即，上述各實施形態並未限 定如上述之構成，只要在實施階段不背離其要旨之範圍內 ，可以實施構成要素之變形並具體化。 .、依據以上說明之本發明，可以大幅減少無需防洪活動 時之防洪活動，並實現有效率之防洪活動。 【圖式簡單說明】 第1圖係本發明第1實施形態之防洪支援裝置實例的 方塊圖。 第2圖係防洪對象區域內之雨量計及雨量計的設置狀 態圖。 第3圖係含設定於資訊設定部之閥値在內之設定資訊 的實例圖。 第4圖係過去之統計資料及閥値的關係實例圖。 第5圖係同實施形態之防洪支援裝置之支援處理部的 處理實例流程圖。 -29- 1242145 (26) 第6圖係畫面顯示用防洪支援資訊之第1例圖。 第7圖係畫面顯示用防洪支援資訊之第2例圖。 第8圖係未使用支援資訊而以有無發佈大雨警報來實 施防洪支援之結果實.例圖。 第9圖係使用支援資訊之防洪支援結果的實例圖。 第1 〇圖係利用評估函數決定閥値時之防洪支援結果 的實例圖。 第1 1圖係本發明第2實施形態之防洪支援裝置之構 成實例方塊圖。 第1 2圖係同實施形態之防洪支援裝置之資料控掘部 的具體實例方塊圖。 第1‘3圖係利用同實施形態之防洪支援裝置執行之處 理的實例流程圖。 第1 4圖係利用電腦實現防洪支援裝置之動作時之構 成實例方塊圖。 [元件符號之說明] 101、102…防洪支援裝置 1、 32···氣象資訊取得部 2、 33…補助資訊取得部 3、 34…補助資訊處理部 4...資訊設定部 5…支援處理部 5 1…氣象判斷部 -30- 1242145 (27) 52··.必要性判斷部 5 3…擬定部 54、35…輸出部 6…顯示裝置 7a、7b…輸入裝置 1 4…雨量計 1 5…水位計 30…資料控掘部 3 1…教師資料取得部 30a...類神經網路 -31(1)… < 2) (3) E = 1/2 (yi (3) -di2} -25- 1242145 (22) In formulas (1) to (4), XI represents input information. M represents class The number of phases of the neural network 3 0a. Y / W represents the activity of the first neuron in phase m. N m represents the number of neurons in phase m. D; represents the danger of each monitoring point (dangerous trunk point). For example 'Teacher poor materials use: input information (X / CM) including rainfall information in the past flood control area, water level information in the sewer trunk line, and meteorological information (data) from the Meteorological Bureau; and Output information including the danger degree (or emergency treatment degree) of each monitoring point at the time when the input information was obtained. More specifically, the input information of the teacher data includes rainfall per unit time in the past flood control target area. Volume information, past single-bit time water level information in the flood prevention area, and "1" when the heavy weather warning was issued by the Meteorological Bureau, and "0" when it was not released. Included in the past Time rainfall is 30 [surface / h], water level is 5.0 [m], and the meteorological bureau issues heavy rain At the time, the input information of the teacher data is (30, 5.0, 1). The output information of the teacher data is the following information. At the time of obtaining the past input information, for each monitoring point, use "1". Information indicating dangers such as reaching a dangerous water level, or water overflow, and "0" representing no danger. At the time of obtaining past input information, if only the monitoring point P2 is in danger, the output information of the teacher data is (0 , 1, ...). When inputting the teacher's information, the data control and mining department 30 will use the teacher's data to implement learning with a neural network 3 0 a, and will repeat the learning until the output -26-! 242145 (23) The square error between the information and the actual neuron output is less than a certain level. The data mining unit 30 will save the composition and weighted data of the neural network 3 oa obtained from learning. After learning, the data mining unit 30 will The meteorological information acquisition unit 32 and the _help information processing unit 34 receive the input information of the current point, and then input the input information to the neural network 3 Oa, and output the risk level of each monitoring point representing the time point. The output layer of the analog neural network 3 0 a will output analog information of the activity level 0 to 1. The data control and excavation department 30 will use the analog information as the danger level (ratio) of each monitoring point. Also, the danger The degree can also be expressed as a percentage. ..; Output, the department 35 will output the flood prevention support information for displaying the danger displayed on the screen. The operator displays the danger at each monitoring point of the display device according to the flood prevention support information. Figure 13 is a flowchart of an example of processing executed by the flood control support device of this embodiment 〇 02. In step T 1, the teacher information acquisition section 31 will refer to the past information recording section. 36 6. Obtain teacher data and provide it to the data mining department 30. In step T2, the data control and excavation department 30 will perform the neural network-like learning according to the teacher data. In step T3a, the meteorological information acquisition unit 32 acquires meteorological information 'from the Meteorological Bureau and performs digitization of the acquired meteorological information. In step T3b, the subsidy information acquisition unit 33 obtains the subsidy information from at least one of the rain gauge and the water level gauge. -27- 1242145 (24) In step T 3 c, the subsidy information processing unit 34 converts the subsidy information into support information. The execution order of steps T3a, T3b, and T3c can be set freely. That is, step T3a may be performed before steps T3b, T3c, or may be performed afterwards, or may be performed in steps T3b, T3c. In step T4, the data mining unit 30 will input the input information including the digitalized meteorological information and support information to the neural network 3 0a. In step T5, the data mining unit 30 obtains output information corresponding to the input information and provides it to the output unit 35. In step T6, the output unit 35 performs control for the purpose of outputting output information to the display device. In this implementation form, it is possible to provide flood control support information based on past experience based on information such as rainfall information and water level information in the flood control support area. Operators can refer to flood support information for highly accurate and rapid staffing activities. In addition, the number of inaccurate flood prevention activities such as on-site inspections where no danger occurs can be greatly reduced. Moreover, in this embodiment, for example, the rainfall information of the past flood control target area, the past water level information in the main sewer line, and the past flood control activity content (data) may be used as input information of the teacher data, and each monitoring point The past risk is used as the output information of the teacher data. At this time, the rainfall information of the current flood prevention target area, the water level information of the waterway trunk below the current point, and the current flood control activity content are used as input information. The current point output can be obtained by using the neural network 3 Oa Information. The implementation of the flood control support devices 101 and 102 in the above-mentioned embodiments -28- 1242145 (25) l ′ can also be realized by a program. As shown in FIG. 14, when the functions of the flood control support devices 10i and 102 are realized by the program 37, the program 37 is recorded on the recording medium 38 and is taken to the computer 39. Secondly, the control units such as C P U included in the computer 39 will perform actions according to the contents of the program 37. Also, in the flood control support devices 1 0 1 and 1 02 of the above-mentioned embodiments, as long as the constituent elements can achieve the same operation, the arrangement can be changed, and the constituent elements can be freely combined, and the constituents can be divided freely. Element, or 'delete several constituent elements. That is, the above-mentioned embodiments are not limited to the structure as described above. As long as the implementation stage does not deviate from the gist thereof, the constituent elements may be modified and embodied. . According to the present invention described above, flood prevention activities when flood prevention activities are not needed can be greatly reduced, and efficient flood prevention activities can be realized. [Brief Description of the Drawings] Fig. 1 is a block diagram of an example of a flood prevention supporting device according to a first embodiment of the present invention. Figure 2 is a diagram of the setting of the rain gauge and the rain gauge in the flood prevention area. Fig. 3 is an example diagram of setting information including a valve set set in the information setting section. Figure 4 is an example of the relationship between the past statistics and the valve valve. Fig. 5 is a flowchart of a processing example of the support processing unit of the flood control support device of the same embodiment. -29- 1242145 (26) Figure 6 is the first example of the flood control support information for screen display. Fig. 7 is a second example of flood control support information for screen display. Figure 8 is an example of the result of implementing flood prevention support with or without issuing a heavy rain alarm without using support information. Figure 9 is an example of the flood support results using support information. Figure 10 is an example of the result of flood control support when the valve is determined using the evaluation function. Fig. 11 is a block diagram showing an example of the structure of a flood control supporting device according to a second embodiment of the present invention. Fig. 12 is a block diagram of a specific example of the data control and excavation department of the flood control support device of the same embodiment. Fig. 1'3 is a flowchart showing an example of processing performed by the flood control supporting device of the same embodiment. Fig. 14 is a block diagram of an example of a structure when a computer is used to implement the operation of the flood control support device. [Explanation of component symbols] 101, 102 ... Flood control support device 1, 32 ... Weather information acquisition section 2, 33 ... Subsidy information acquisition section 3, 34 ... Subsidy information processing section 4 ... Information setting section 5 ... Support processing Unit 5 1… Meteorological judgment unit -30-1242145 (27) 52 · .. Necessity judgment unit 5 3… Designation unit 54, 35… Output unit 6… Display device 7a, 7b… Input device 1 4… Rain gauge 1 5 … Water level meter 30… Data control and excavation unit 3 1… Teacher data acquisition unit 30a ... Class neural network-31