201113724 六、發明說明： 【發明所屬之技術領域】 本發明關於搜索技術領域，特別關於一種圖像資料庫 中的圖像更新方法、伺服器及系統。 【先前技術】 目前的圖像搜索技術，包含兩層含義，一層含義是指 搜索出圖像，另一層含義是指依據圖像內容進行的搜索。 比較好理解，前一層含義的圖像搜索，可以通過搜索詞搜 索出圖像資料庫中命中檔案名的圖像。而後一層含義的圖 像搜索，則是以搜索請求中的圖像爲基礎，搜索出圖像資 料庫中與搜索請求中的圖像內容相類似的圖像。本篇討論 的圖像資料庫中的圖像更新方法及伺服器，適用於後一層 含義的圖像搜索。 本篇涉及的圖像搜索，現有技術的過程可以如圖1中 所示，描述爲如下步驟： S110:搜尋引擎接收搜索請求。該搜索請求中包括要 求捜索的圖像。 S 1 20 :搜尋引擎獲取搜索請求中要求搜索的圖像的特 徵。具體的，搜尋引擎可以通過一些特定的圖像計算方法 提取出要求搜索圖像的特徵，例如可以是其中包含物體的 形狀特徵。 S130:搜尋引擎在圖像資料庫中搜索與提取的特徵最 接近的圖像，返回搜索結果。 -5- 201113724 上述過程中，搜尋引擎需要在圖像資料庫中 ，即找出圖像資料庫中與提取特徵接近的圖像的 然，圖像資料庫中，要預先計算出所保存的圖像 以便於接到搜索請求時可以直接將這些特徵與前 的搜索請求中的圖像特徵作比較。 圖像資料庫中的圖像會產生不斷的更新，以 更多的圖像，從而有能力爲搜索提供更豐富的結 現有的圖像資料庫中的圖像更新，是定時將 與新增的圖像一併更新到圖像資料庫中，並重新 。上述更新過程一般被安排在每天作爲搜尋引擎 最閒的時段做這一處理，例如浚晨2點至3點， 間段內，例如將1 000多萬的歷史圖像與新增的 像一倂更新到圖像資料庫中。 在對現有技術的硏究和實踐過程中，發明人 技術中存在以下問題： 現有技術中這種圖像的更新方法，需要將歷 新增的圖像一倂更新，由於歷史圖像數量較大， 的更新，需要的時間較長。 【發明內容】 本申請案實施例的目的是提供一種圖像資料 像更新方法、伺服器及系統，以降低圖像更新過 時間。 爲解決上述技術問題，本申請案實施例提供 進行搜索 特徵。當 的特徵， 述提取出 收集提供 累。 歷史圖像 建立索引 的伺服器 在這一時 2 0 0萬圖 發現現有 史圖像與 進行這樣 庫中的圖 程所需的 —種圖像 -6 - 201113724 資料庫中的圖像更新方法及伺服器是這樣實現的： 一種圖像資料庫中的圖像更新方法，包括： 獲取發生更新的圖像； 提取更新的圖像的特徵値並儲存； 將提取的更新圖像的特徵値關聯於搜尋引擎。 較佳地，所述方法中，所述獲取發生更新的圖像，包 括： 通過圖像更新表中更新的圖像ID獲得更新的圖像。 較佳地，所述方法中，所述獲取發生更新的圖像，包 括： 在記錄圖像的圖像資料庫中設置檢測圖像發生變化的 觸發器，利用所述觸發器檢測到圖像資料庫中更新圖像時 伴隨的更新命令發生時，記錄發生更新的圖像的ID。 較佳地’所述方法中，所述提取更新的圖像的特徵値 ，包括： 以預定週期比較圖像更新表與圖像特徵表相同圖像 ID對應的修改時間； 如果圖像特徵表中的修改時間不存在，則提取該圖像 ID對應圖像的特徵値並記錄於圖像特徵表中，並記錄當 前時刻作爲修改時間。 較佳地’所述方法中，所述提取更新的圖像的特徵値 ，包括： 以預定週期比較圖像更新表與圖像特徵表相同圖像 ID對應的修改時間； 201113724 如果圖像特徵表中的修改時間早於圖像更新表中的修 改時間，則提取該圖像ID對應圖像的特徵値並更新於圖 像特徵表中，並更新當前時刻作爲修改時間。 較佳地，所述方法中，對於分散式搜尋引擎，還包括 將提取的更新圖像特徵値分發給分散式搜尋引擎。 —種更新圖像資料庫中圖像的伺服器，包括： 獲取單元，用於獲取發生更新的圖像； 提取單元，用於提取更新的圖像的特徵値並儲存； 關聯單元，用於將提取的更新圖像特徵値關聯於搜尋 引擎。 較佳地，所述伺服器中，所述獲取單元通過圖像更新 表中更新的圖像ID獲得更新的圖像。 較佳地，所述伺服器中，所述獲取單元中包括圖像資 料庫，觸發器，記錄單元，其中， 圖像資料庫，用於記錄圖像： 觸發器，用於檢測圖像資料庫中的圖像是否發生變化 > 記錄單元，用於在觸發器檢測到圖像資料庫中更新圖 像時伴隨的更新命令發生時，記錄發生更新的圖像的ID 〇 較佳地，所述伺服器中，所述提取單元包括： 比較單元，用於以預定週期比較圖像更新表與圖像特 徵表相同圖像ID對應的修改時間； -8- 201113724 第一記錄單元，當圖像特徵表中的修改時間不存在時 ，用於提取該圖像ID對應圖像的特徵値並記錄於圖像特 徵表中，並記錄當前時刻作爲修改時間。 較佳地，所述伺服器中，所述提取單元包括： 比較單元，用於以預定週期比較圖像更新表與圖像特 徵表相同圖像ID對應的修改時間； 第二記錄單元，如果圖像特徵表中的修改時間早於圖 像更新表中的修改時間，則提取該圖像ID對應圖像的特 徵値並更新於圖像特徵表中，並更新當前時刻作爲修改時 較佳地，所述伺服器中，對於分散式搜尋引擎，所述 伺服器還包括： 分發單元’用於將提取的更新圖像的特徵値分發給分 散式搜尋引擎。 一種更新圖像資料庫中圖像的系統，伺服器以及搜尋 引擎，其中， 搜尋引擎用於與網站伺服器提取的更新圖像的特徵値 關聯。 一種更新圖像資料庫中圖像的系統，包括伺服器以及 搜尋引擎，其中， 伺服器還用於接收並儲存用戶端上傳的圖像； 搜尋引擎’用於與所述伺服器提取的更新圖像的特徵 値相關聯。 一種更新圖像資料庫中圖像的系統，包括伺服器以及 -9- 201113724 網際網路圖像抓取子系統、搜尋引擎，其中， 網際網路圖像抓取子系統，用於抓取網際網路上的圖 像並傳輸至伺服器； 伺服器還用於接收並儲存所述網際網路圖像抓取子系 統傳輸的圖像； 搜尋引擎，用於與所述伺服器提取的更新圖像的特徵 値相關聯。由以上本申請案實施例提供的技術方案可見， 獲取發生更新的圖像，提取更新的圖像的特徵値並儲存， 將提取的更新圖像的特徵値關聯於搜尋引擎，由於整個過 程只涉及發生圖像更新的圖像，而不對未發生更新的圖像 進行處理，更不是對涉及整個圖像資料庫中的圖像的處理 ’因此，相對於現有技術，可以大大降低提取更新的圖像 的特徵値所需的時間，並大大降低將特徵値關聯於搜尋引 擎所需的時間，進而可以大大降低整個更新過程所需的時 間。 【實施方式】 本申請案實施例提供一種圖像資料庫中的圖像更新方 法及伺服器。 爲了使本技術領域的人員更好地理解本申請案中的技 術方案，下面將結合本申請案實施例中的附圖’對本申請 案實施例中的技術方案進行清楚、完整地描述’顯然，所 描述的實施例僅僅是本申請案一部分實施例，而不是全部 的實施例。基於本申請案中的實施例’本領域普通技術人 -10- 201113724 員在沒有做出創造性勞動前提下所獲得的所有其他實施例 ，都應當屬於本發明保護的範圍。 圖2示出了本申請案圖像更新方法實施例的流程圖， 如圖2所示，包括： S210:獲取發生更新的圖像^ 首先需要獲取發生更新的圖像，如獲取新增的圖像， 或者如獲取發生修改等變化的圖像，而非所有的圖像，這 是保證後續僅對發生更新的圖像提取特徵値並在搜尋引擎 中建立索引的基礎。 具體的，可以通過圖像更新表中更新的圖像ID獲得 更新的圖像。 負責更新圖像的系統中，存在圖像更新表，該表中記 錄發生更新的圖像ID。該圖像更新表中記錄的圖像ID， 可以是人工記錄的，也可以是其他方式記錄得到，如從其 他儲存圖像的系統或資料庫中導入而得到。 以下給出一種圖像更新表中自動記錄發生更新的圖像 ID的方式。可以在記錄所有圖像的圖像資料庫中設置— 個檢測圖像發生變化的觸發器（Trigger )。一般新增加的 圖像都會首先增加到這個圖像資料庫中，當然，該圖像資 料庫中其他對圖像的更改操作，如對圖像的修改等，也會 導致這個圖像資料庫中對圖像發生變化進行記載。圖像資 料庫中，記錄的每個圖像都有唯一的ID進行標識。 如果是圖像資料庫中新增的圖像，則會賦予—個新的 唯一 ID。新增的圖像，可以是用戶終端上傳的圖像，也 -11 - 201113724 可以是通過搜索技術中的爬蟲系統從網際網路中抓取的圖 像。如果是用戶上傳的新的圖像，在用戶上傳過程中，可 以選擇該圖像的上傳方式是新增，則圖像資料庫中記錄該 新增圖像的過程中’如前所述，賦予該新增圖像一個唯— ID。如果是爬蟲系統抓取的圖像’則根據該抓取圖像來源 的鏈結位址’或根據該抓取圖像存在的來原始目錄以及檔 案名 '檔格式等’通過與圖像資料庫中已存在圖像相應的 來源鏈結位址進行比較’或相應地與圖像資料庫中已存在 圖像所在的來原始目錄以及檔案名、檔格式等進行比較， 如果沒有相同的’從而可以得出不是同一圖像，進而可以 判斷該圖像是新增圖像，則如前所述，則圖像資料庫中記 錄該新增圖像的過程中，賦予該新增圖像一個唯一 ID。 如果是圖像資料庫中發生修改等更新操作的圖像，則 保持更新前圖像的ID。發生更新操作的圖像，類似的， 可以是用戶終端上傳的圖像，也可以是通過搜索技術中的 爬蟲系統從網際網路中抓取的圖像。如果是用戶終端上傳 的圖像，在用戶上傳過程中，可以選擇該圖像的上傳方式 是修改等方式進行更新，則圖像資料庫記錄該更新圖像的 過程中，保持該圖像的ID與修改前一致。如果是爬蟲系 統抓取的圖像，則根據抓取圖像來源的鏈結位址，或根據 抓取圖像存在的來原始目錄以及檔案名、檔格式等，通過 與圖像資料庫中已存在圖像相應的來源位址進行比較，或 相應地與圖像資料庫中已存在圖像所在的來原始目錄以及 檔案名、檔格式等進行比較’如果有相同的’則可以得出 -12- 201113724 是同一圖像’進而可以判斷該圖像是更新的圖像，則如前 所述’則圖像資料庫中記錄該新增圖像的過程中，保持 ID與更新前相同。 無論是新增的圖像’還是發生修改等更新操作的圖像 ’在圖像資料庫記錄過程中，一般都會伴隨發生資料庫中 的Update命令。則前述提到的增加的觸發器，可以檢測 Up d at e命令的發生’如果檢測到，即說明有圖像資料庫的 更新。從而，圖像更新表中將記錄圖像資料庫中更新的圖 像ID，進而，可以通過檢測圖像更新表中圖像id得到更 新的圖像。 S 220 :提取更新的圖像的特徵値並儲存。 更新的圖像，可以提取出其特徵値。提取出的特徵値 可以爲一定維度的向量，如（Ρ1，Ρ2，·..，Ρη)，這個η 維向量，例如可以是用來描述圖像中物體的形狀特徵。具 體的提取方式，可以採取現有的圖像特徵提取方式，如採 用Hough變換等方式的圖像特徵提取，在此不再介紹。另 外，以下簡單介紹一種較優的圖像特徵提取的方式，可以 包括下面步驟= A 1 :從原始圖像中樞出所含物體的圖像； A2:將所述樞出的圖像用單一顏色爲背景塡充邊界’ 並使得塡充後的圖像成爲最小正方形； A3:將正方形圖像全圖等比縮放爲第一預定大小的圖 像，將縮放後的圖像分割爲第二預定大小的子圖像塊； A4:分別計算子圖像塊水準、豎直、正45°、負45° -13- 201113724 方向上相鄰圖元的亮度導數，將分別在四個方向導數極値 點的個數、以及位於子圖像塊四個邊界上極値點的總個數 作爲該子圖像塊的特徵向量； A5:將所有子圖像塊的特徵向量作爲原始圖像的特徵 向量。 上述圖像特徵提取方式，主要利用了圖像中顯示的物 體其邊緣部分與周圍的背景等的圖元亮度差異來找出物體 的邊緣’也就可以提取得到圖像中物體的形狀特徵，從而 可以作爲該圖像提取的特徵。 可以設置一圖像特徵表，用於儲存圖像的特徵値。提 取出的更新圖像的特徵値，被儲存於該表中。 提取操作的執行，可以按照一定規則設定。以下給出 一種執行規則’在介紹該規則前，先給出圖像更新表和圖 像特徵表的具體形式示例： 圖像ID 修改時間 A1 2009-1-1 18:00:00 A2 2009-1-1 17:00：〇〇 • · · • · · 表1.圖像更新表 圖像ID 特徵値 修改時間 A1 (Pll，P12，...，pm) 2009-1-1 17:00:00 A2 (P21，P22，…，p2n) 2009-1-1 18:00:00 ♦.. 表2.圖像特徵表 -14 - 201113724 則提取規則，可以如下： 以預定週期比較圖像更新表與圖像特徵表相同圖像 ID對應的修改時間， 如果圖像特徵表中的修改時間不存在，則提取該圖像 ID對應圖像的特徵値並記錄於圖像特徵表中，並記錄當 前時刻作爲修改時間； 如果圖像特徵表中的修改時間早於圖像更新表中的修 改時間，則提取該圖像ID對應圖像的特徵値並更新於圖 像特徵表中’並更新當前時刻作爲修改時間。 另外’如果圖像特徵表中的修改時間不早於圖像更新 表中的修改時間，則可以不動作。 對於圖像資料庫中新增的圖像，圖像更新表中會記錄 該新增圖像的ID，但是圖像特徵表中沒有該新增圖像的 ID。這種情況，則按照上述提取規則進行提取之前，還可 以包括：對於圖像更新表中存在的圖像ID而在圖像特徵 表中不存在的情況，將所述圖像ID記錄於圖像特徵表中 。這樣’在提取特徵步驟之前’圖像特徵表中將僅包含圖 像更新表中記載的新增圖像的ID，但是，並沒有包含該 新增圖像的修改時間等內容。進而，在後續過程，可以按 照上述提取規則提取新增圖像的特徵並記錄於圖像特徵表 中〇 則按照上述規則’表1的圖像更新表中，圖像ID爲 A1的修改時間爲2 0 0 9年1月1日1 8點〇 1分整，而表2 圖像特徵表中’圖像ID爲A1的修改時間爲2009年1月 -15- 201113724 1日17點整，比表1中的時間要早，說明id爲A1的圖 像發生更新後，圖像特徵表中並沒有提取過該圖像的特徵 。因此，根據上述規則，將提取A1的特徵値，並更新於 圖像特徵表中，並更新A1的修改時間爲當前時刻，例如 爲2 009年1月1日18點10分整，則更新後的圖像特徵 表可以如下面表3所示： 圖像ID 特徵値 修改時間 A1 (PI 1，，P12,，…，Pin，） 2009-1-1 18:10:00 A2 (P21 ， P22 ，…，P2n) 2009-1-1 18:00:00 … * · * • . · 表3·更新後的圖像特徵表 上述規則，保證了圖像特徵表中及時記錄更新的圖像 ID、最後記錄的修改時間及其更新後的特徵値。 S 2 3 0 :將提取的更新圖像特徵値關聯於搜尋引擎。 具體的，可以包括爲更新圖像提取的特徵値在搜尋引 擎中建立索引。 實際上，該步驟的目的是將更新的圖像build進搜 尋引擎。BUILD過程，包括將要被搜索的圖像建立索引， 放到記憶體等過程。 特別的’每個圖像的特徵値在索引上建立一個陣列。 這是由於圖像特徵値包括一個向量，這個向量中包含若干 値，因此需要採用陣列來存放。 對於分散式搜尋引擎，本方法實施例還可以包括： -16- 201113724 s 240 :將提取的更新圖像特徵値分發給分散式搜尋引 擎。 由以上實施例可見，獲取發生更新的圖像，提取更新 的圖像的特徵値並儲存，將提取的更新圖像特徵値關聯於 搜尋引擎，由於整個過程只涉及發生更新的圖像，而不對 未發生更新的圖像進行處理，更不是對涉及整個圖像資料 庫中的圖像的處理，因此，相對於現有技術，可以大大降 低提取更新的圖像的特徵値所需的時間，並大大降低將特 徵値關聯於搜尋引擎所需的時間，進而大大降低整個更新 過程所需的時間。 以下介紹本申請案一更新圖像資料庫中圖像的伺服器 實施例，圖3示出了該伺服器實施例的方塊圖，如圖，包 括： 獲取單元31，用於獲取發生更新的圖像； 提取單兀32，用於提取更新的圖像的特徵値並儲存； 關聯單元3 3，用於將提取的更新圖像特徵値關聯於搜 尋引擎。 較佳地，所述伺服器中，所述獲取單元3 1可以通過 圖像更新表中更新的圖像ID獲得更新的圖像。其中，所 述圖像更新表可以儲存於伺服器中的一儲存單元內，當然 ’也可以儲存於伺服器之外其他設備上的一儲存單元內， 這裏並不限定。 較佳地’所述伺服器中，所述獲取單元31中具體可 以包括觸發器311’記錄單元312，如圖4所示，其中， -17- 201113724 觸發器311’用於檢測圖像資料庫中的圖像是否發生 變化； 記錄單元3 1 2，用於在觸發器檢測到圖像資料庫中更 新圖像時伴隨的更新命令發生時，記錄發生更新的圖像的 ID。 需要說明的是，上述圖4中所示的組成和連接關係只 是本申請伺服器的一種示例。除此之外，所述資料庫和觸 發器也可以位於伺服器之外的其他設備之上，這主要取決 於哪個設備負責承擔圖像資料庫功能。 較佳地，所述伺服器中，如圖5所示，所述提取單元 3 2可以包括： 比較單元321，用於以預定週期比較圖像更新表與圖 像特徵表相同圖像ID對應的修改時間； 第一記錄單元322，當圖像特徵表中的修改時間不存 在時，用於提取該圖像ID對應圖像的特徵値並記錄於圖 像特徵表中，並記錄當前時刻作爲修改時間。 較佳地，所述伺服器中，如圖6所示，所述提取單元 3 2包括： 比較單元321，用於以預定週期比較更新表與圖像特 徵表相同圖像ID對應的修改時間； 第二記錄單元3 23，如果圖像特徵表中的修改時間早 於圖像更新表中的修改時間，則提取該圖像ID對應圖像 的特徵値並更新於圖像特徵表中，並更新當前時刻作爲修 改時間。 -18- 201113724 較佳地’所述伺服器中’如圖7所示，對於分散式搜 尋引擎’所述伺服器還包括： 分發單元34，用於將提取的更新圖像特徵値分發給分 散式搜尋引擎。 爲了描述的方便’描述以上裝置時以功能分爲各種單 元分別描述。當然’在實施本發明時可以把各單元的功能 在同一個或多個軟體和/或硬體中實現。 以下介紹本申請的另一種更新圖像資料庫中圖像的系 統貫施例’該系統實施例可以如圖8所示，包括伺服器8 1 和搜尋引擎82，其中， 伺服器81可以爲上述圖3至圖7中任一種伺服器； 搜尋引擎82 ’用於與伺服器81提取的更新圖像的特 徵値關聯。 以下介紹本申請的另一種更新圖像資料庫中圖像的系 統’可以如圖9所示’包括伺服器9 1以及用戶終端9 3、 搜尋引擎92，其中， 用戶終端9 3，用於通過網際網路上傳圖像至伺服器； 伺服器91可以爲上述圖3至圖7中任—種伺服器， 並還接收並儲存用戶終端93上傳的圖像； 搜尋引擎92，用於與所述伺服器91提取的更新圖像 的特徵値相關聯。 以下介紹本申請的另一種更新圖像資料庫中圖像的系 統’可以如圖1 0所示’包括伺服器〗〇 i以及網際網路圖 像抓取子系統103、搜尋引擎1〇2，其中， -19- 201113724 網際網路圖像抓取子系統103，用於抓取網際網路上 的圖像並傳輸至伺服器1〇1:如前所述，網際網路圖像抓 取子系統103，可以是利用搜索技術中的爬蟲系統抓取的 網際網路中的圖像。 伺服器1 〇 1可以爲上述圖3至圖7中任一種伺服器， 並還接收並儲存所述網際網路圖像抓取子系統103傳輸的 圖像， 搜尋引擎102，用於與所述伺服器101提取的更新圖 像的特徵値相關聯。 如前所述，網際網路圖像抓取子系統1 〇3，可以是利 用搜索技術中的爬蟲系統抓取的網際網路中的圖像。 需要說明的是，上述圖10中所示的組成和連接關係 只是本申請案系統的一種示例。除此之外，圖10中的網 際網路圖像抓取子系統1 03還可以與伺服器1 0 1或搜尋引 擎1〇2集成在一起。 通過以上的實施方式的描述可知，本領域的技術人員 可以清楚地瞭解到本申請案可借助軟體加必需的通用硬體 平臺的方式來實現。基於這樣的理解，本申請案的技術方 案本質上或者說對現有技術做出貢獻的部分可以以軟體產 品的形式體現出來，該電腦軟體產品可以儲存在儲存媒體 中，如ROM/RAM、磁碟、光碟等，包括若干指令用以使 得一台電腦設備（可以是個人電腦，伺服器，或者網路設 « 備等）執行本申請案各個實施例或者實施例的某些部分所 述的方法。 -20- 201113724 本申請案中的各個實施例均採用遞進的方式描述，各 個實施例之間相同相似的部分互相參見即可’每個實施例 重點說明的都是與其他實施例的不同之處。尤其，對於系 統實施例而言，由於其基本相似於方法實施例，所以描述 的比較簡單，相關之處參見方法實施例的部分說明即可。 本申請可用於眾多通用或專用的什算系統環境或配置 中。例如：個人電腦、伺服器電腦、手持設備或可檇式設 備、平板型設備、多處理器系統、基於微處理器的系統、 機頂盒、可編程的消費電子設備、網路PC、小型電腦、 大型電腦、包括以上任何系統或設備的分散式計算環境等 等。 本申請案可以在由電腦執行的電腦可執行指令的一般 上下文中描述，例如程式模組。一般地，程式模組包括執 行特定任務或實現特定抽象資料類型的常式、程式 '物件 、元件、資料結構等等。也可以在分散式計算環境中實踐 本發明，在這些分散式計算環境中，由通過通信網路而被 連接的遠端處理設備來執行任務。在分散式計算環境中， 程式模組可以位於包括儲存設備在內的本地和遠端電腦儲 存媒體中。 雖然通過實施例描繪了本發明，本領域普通技術人員 知道’本發明有許多變形和變化而不脫離本發明的精神， 希望所附的權利要求包括這些變形和變化而不脫離本發明 的精神。 -21 - 201113724 【圖式簡單說明】 爲了更清楚地說明本申請案實施例或現有技術中的技 術方案’下面將對實施例或現有技術描述中所需要使用的 附圖作簡單地介紹，顯而易見地，下面描述中的附圖僅僅 是本申請中記載的一些實施例，對於本領域普通技術人員 來講’在不付出創造性勞動性的前提下，還可以根據這些 附圖獲得其他的附圖。 圖1爲現有技術中圖像搜索方案的流程圖； 圖2爲本申請案圖像更新方法實施例的流程圖； 圖3爲本申請案一伺服器實施例的方塊圖； 圖4爲本申請案另一伺服器實施例的方塊圖； 圖5爲本申請案另一伺服器實施例的方塊圖； 圖6爲本申請案另一伺服器實施例的方塊圖； 圖7爲本申請案另一伺服器實施例的方塊圖： 圖8爲本申請案另一系統實施例的方塊圖； 圖9爲本申請案另一系統實施例的方塊圖； 圖10爲本申請案另一系統實施例的方塊圖。 【主要元件符號說明】 31 :獲取單元 32 :提取單元 33 :關聯單元 31 1 :觸發器 3 1 2 :記錄單元 -22- 201113724 3 2 1 :比較單元 3 22 :第一記錄單元 3 2 3 :第二記錄單元 3 4 :分發單元 8 1 :伺服器 82 :搜尋引擎 9 1 :伺服器 92 :搜尋引擎 93 :用戶終端 1 〇 1 :伺服器 102 :搜尋引擎 1 03 :網際網路圖像抓取子系統 -23-201113724 VI. Description of the Invention: [Technical Field] The present invention relates to the field of search technology, and more particularly to an image update method, server and system in an image database. [Prior Art] The current image search technology includes two layers of meaning, one layer means searching for an image, and the other layer means searching based on image content. It is better to understand, the image search of the previous layer meaning, you can search for the image of the hit file name in the image database by searching for the word. The image search of the latter layer is based on the image in the search request, and searches for an image in the image database similar to the image content in the search request. The image update method and server in the image database discussed in this article are applicable to the image search of the meaning of the latter layer. For the image search involved in this section, the prior art process can be described as follows in FIG. 1 : S110: The search engine receives the search request. The search request includes an image requesting a search. S 1 20: The search engine acquires the characteristics of the image requested to be searched in the search request. Specifically, the search engine may extract features of the image to be searched by some specific image calculation methods, for example, may include shape features of the object. S130: The search engine searches the image database for the image closest to the extracted feature, and returns the search result. -5- 201113724 In the above process, the search engine needs to find the image in the image database that is close to the extracted feature in the image database. In the image database, the saved image is pre-calculated. These features can be directly compared to the image features in the previous search request in order to receive the search request. The images in the image library will generate constant updates to more images, thus the ability to provide a richer image for the search. The image updates in the existing image database are timed and added. The image is updated to the image database and re-created. The above update process is generally arranged to do this process as the free time of the search engine every day, for example, from 2 am to 3 am, in the interval, for example, the historical image of more than 10 million is added to the new image. Update to the image library. In the research and practice of the prior art, the following problems exist in the inventor's technology: In the prior art, such an image updating method needs to update the newly added image, because the number of historical images is large. , the update takes longer. SUMMARY OF THE INVENTION An object of an embodiment of the present application is to provide an image data update method, a server, and a system to reduce image update time. In order to solve the above technical problems, embodiments of the present application provide for performing search features. When the characteristics are extracted, the collection is provided to provide tiring. The historical image indexing server at this time finds the existing history image and the image required to perform the drawing in such a library. The image updating method and servo in the database -6 - 201113724 database The device is implemented as follows: an image updating method in an image database, comprising: acquiring an image that is updated; extracting features of the updated image and storing; and correlating the extracted features of the updated image with the search engine. Preferably, in the method, the acquiring the updated image comprises: obtaining the updated image by using the updated image ID in the image update table. Preferably, in the method, the acquiring the updated image comprises: setting a trigger for detecting a change of the image in the image database of the recorded image, and detecting the image data by using the trigger When the update command accompanying the update of the image in the library occurs, the ID of the image in which the update occurred is recorded. Preferably, in the method, the extracting the feature of the updated image comprises: comparing, at a predetermined period, a modification time corresponding to the same image ID of the image update table and the image feature table; if the image feature table is If the modification time does not exist, the feature 値 of the image corresponding to the image ID is extracted and recorded in the image feature table, and the current time is recorded as the modification time. Preferably, in the method, the extracting the feature of the updated image comprises: comparing the modification time corresponding to the same image ID of the image update table and the image feature table by a predetermined period; 201113724 if the image feature table The modification time in the image is earlier than the modification time in the image update table, and the feature 値 of the image corresponding to the image ID is extracted and updated in the image feature table, and the current time is updated as the modification time. Preferably, in the method, for the distributed search engine, the method further includes distributing the extracted updated image feature to the distributed search engine. a server for updating an image in an image database, comprising: an obtaining unit for acquiring an image in which an update occurs; an extracting unit for extracting a feature of the updated image and storing; an associating unit, configured to The extracted updated image features are associated with the search engine. Preferably, in the server, the acquisition unit obtains the updated image by the updated image ID in the image update table. Preferably, in the server, the acquiring unit includes an image database, a trigger, and a recording unit, wherein the image database is used for recording an image: a trigger for detecting an image database Whether the image in the image changes > a recording unit for recording an ID of the image in which the update occurs when the trigger is accompanied by an update command accompanying the update of the image in the image database 〇 preferably, said In the server, the extracting unit includes: a comparing unit, configured to compare the modification time corresponding to the same image ID of the image update table and the image feature table by a predetermined period; -8- 201113724, the first recording unit, when the image feature When the modification time in the table does not exist, the feature 値 for extracting the image corresponding to the image ID is recorded and recorded in the image feature table, and the current time is recorded as the modification time. Preferably, in the server, the extracting unit comprises: a comparing unit, configured to compare, at a predetermined period, a modification time corresponding to the same image ID of the image update table and the image feature table; the second recording unit, if If the modification time in the feature table is earlier than the modification time in the image update table, extracting the feature 値 of the image corresponding to the image ID and updating it in the image feature table, and updating the current time as a modification, preferably, In the server, for the distributed search engine, the server further includes: a distribution unit 'for distributing the feature 値 of the extracted update image to the distributed search engine. A system for updating images in an image database, a server, and a search engine, wherein the search engine is associated with a feature of the updated image extracted by the web server. A system for updating an image in an image database, comprising a server and a search engine, wherein the server is further configured to receive and store an image uploaded by the client; and the search engine is configured to update the image extracted by the server The characteristics of the image are associated. A system for updating images in an image database, including a server and an Internet image capture subsystem, a search engine, wherein the Internet image capture subsystem is used to capture the Internet An image on the network is transmitted to the server; the server is further configured to receive and store an image transmitted by the internet image capture subsystem; a search engine for updating the image extracted with the server The characteristics are associated with each other. The technical solution provided by the embodiment of the present application can be used to obtain an updated image, extract the feature of the updated image, and store the feature of the extracted updated image with the search engine, since the entire process only involves An image-updated image occurs without processing the image that has not been updated, and is not a process involving the image in the entire image database. Therefore, the updated image can be greatly reduced compared to the prior art. The characteristics of the time required, and greatly reduce the time required to correlate the feature to the search engine, which can greatly reduce the time required for the entire update process. [Embodiment] An embodiment of the present application provides an image updating method and a server in an image database. In order to enable those skilled in the art to better understand the technical solutions in the present application, the technical solutions in the embodiments of the present application will be clearly and completely described in the following with the accompanying drawings in the embodiments of the present application. The described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present application, in the absence of creative work, should be within the scope of the present invention. FIG. 2 is a flowchart of an embodiment of an image updating method of the present application. As shown in FIG. 2, the method includes the following steps: S210: Acquire an image that is updated. ^ First, an image that needs to be updated is acquired, such as obtaining a new image. For example, or if you want to obtain an image with changes such as changes, not all images, this is the basis for ensuring that only the extracted images are extracted and indexed in the search engine. Specifically, the updated image can be obtained by the updated image ID in the image update table. In the system responsible for updating the image, there is an image update table in which the image ID in which the update occurred is recorded. The image ID recorded in the image update table may be manually recorded, or may be recorded in other manners, such as from other systems or databases for storing images. The manner in which the updated image ID is automatically recorded in an image update table is given below. You can set up a trigger (Trigger) that detects changes in the image in the image library where all images are recorded. Generally, the newly added image will be added to the image database first. Of course, other changes to the image in the image database, such as the modification of the image, will also result in the image database. The change in image is described. In the image library, each image recorded has a unique ID for identification. If it is a new image in the image library, it will be given a new unique ID. The newly added image can be the image uploaded by the user terminal, and -11 - 201113724 can be an image captured from the Internet by the crawler system in the search technology. If it is a new image uploaded by the user, during the user uploading process, the uploading mode of the image may be selected to be added, and the process of recording the newly added image in the image database is as described above. The new image is a unique ID. If the image captured by the crawler system is 'according to the link address of the captured image source' or the original directory according to the captured image and the file name 'file format, etc.' through the image database The corresponding source link address of the existing image is compared for 'or correspondingly compared with the original directory where the existing image in the image database is located, as well as the file name, file format, etc., if there is no the same 'and thus If it is not the same image, and then it can be judged that the image is a new image, as described above, in the process of recording the newly added image in the image database, a unique ID is given to the newly added image. . If an image of an update operation such as a modification occurs in the image database, the ID of the image before the update is maintained. The image in which the update operation occurs, similarly, may be an image uploaded by the user terminal, or may be an image captured from the Internet by a crawler system in the search technology. If the image is uploaded by the user terminal, during the user uploading process, the uploading mode of the image may be selected to be updated, and the image database keeps the ID of the image during the process of recording the updated image. Consistent with the changes. If it is an image captured by the crawling system, according to the link address of the captured image source, or according to the existing image of the captured image, the original directory and the file name, file format, etc., through the image database There is a corresponding source address of the image for comparison, or correspondingly compared with the original directory where the image already exists in the image database, as well as the file name, file format, etc. 'If there is the same', then -12 - 201113724 is the same image' and can be judged that the image is an updated image. As described above, the ID is kept the same as the previous one in the process of recording the newly added image in the image database. Whether it is a new image or an image of an update operation such as a modification, the update command in the database is generally accompanied by the image database recording process. Then, the increased trigger mentioned above can detect the occurrence of the Up d at e command. If detected, it indicates that there is an update of the image database. Thereby, the image ID updated in the image database is recorded in the image update table, and further, the updated image can be obtained by detecting the image id in the image update table. S220: Extract the feature of the updated image and store it. The updated image can be extracted from its characteristics. The extracted feature 値 can be a vector of a certain dimension, such as (Ρ1, Ρ2, . . . , Ρη), which can be, for example, a shape feature used to describe an object in an image. The specific extraction method can adopt the existing image feature extraction method, such as image feature extraction using Hough transform, and will not be introduced here. In addition, the following briefly introduces a better image feature extraction method, which may include the following steps: A1: pivoting the image of the contained object from the original image; A2: using the pivoted image in a single color Filling the boundary for the background' and making the expanded image the smallest square; A3: scaling the square image full image to the image of the first predetermined size, and dividing the scaled image into a second predetermined size Sub-image block; A4: Calculate the brightness derivative of the neighboring primitives in the sub-image block level, vertical, positive 45°, negative 45° -13-201113724 directions, respectively, and the derivative points in the four directions The number of the number, and the total number of extreme points on the four boundaries of the sub-image block as the feature vector of the sub-image block; A5: The feature vector of all the sub-image blocks is used as the feature vector of the original image. The image feature extraction method described above mainly utilizes the difference in brightness of the object between the edge portion of the object displayed in the image and the surrounding background to find the edge of the object, thereby extracting the shape feature of the object in the image, thereby Can be used as a feature of this image extraction. An image feature table can be set for storing the features of the image. The characteristics of the extracted updated image are stored in the table. The execution of the extraction operation can be set according to certain rules. An execution rule is given below. Before introducing the rule, an example of the specific form of the image update table and the image feature table is given: Image ID modification time A1 2009-1-1 18:00:00 A2 2009-1 -1 17:00:〇〇• · · • · · Table 1. Image update table image ID characteristics 値 modification time A1 (Pll, P12,..., pm) 2009-1-1 17:00:00 A2 (P21, P22,..., p2n) 2009-1-1 18:00:00 ♦.. Table 2. Image Feature Table-14 - 201113724 The extraction rules can be as follows: Compare the image update table with a predetermined period and The modification time corresponding to the same image ID of the image feature table, if the modification time in the image feature table does not exist, extract the feature 値 of the image corresponding to the image ID and record it in the image feature table, and record the current time As the modification time; if the modification time in the image feature table is earlier than the modification time in the image update table, extracting the feature 値 of the image corresponding to the image ID and updating it in the image feature table 'and updating the current time as Change the time. In addition, if the modification time in the image feature table is not earlier than the modification time in the image update table, it may not operate. For the newly added image in the image database, the ID of the newly added image is recorded in the image update table, but the ID of the newly added image is not in the image feature table. In this case, before the extraction according to the extraction rule, the method may further include: recording the image ID in the image if the image ID is not present in the image feature table for the image ID existing in the image update table. In the feature table. Thus, the image feature table will contain only the ID of the newly added image described in the image update table before the feature extraction step, but does not include the modification time and the like of the new image. Further, in the subsequent process, the feature of the newly added image may be extracted according to the above extraction rule and recorded in the image feature table, and the modification time of the image ID A1 in the image update table according to the above rule 'Table 1 is January 1, 1st, 1st, 1st, 1st, 1st, and the change of the image ID of A1 in the image feature table of Table 2 is January 17th, 2009-201113724, 1st, 17th, The time in Table 1 is earlier, and after the image with the id of A1 is updated, the features of the image are not extracted in the image feature table. Therefore, according to the above rule, the feature A of A1 will be extracted and updated in the image feature table, and the modification time of A1 is updated to the current time, for example, at 18:10 on January 1, 009, after the update The image feature table can be as shown in Table 3 below: Image ID Feature 値 Modification time A1 (PI 1, P12,,..., Pin,) 2009-1-1 18:10:00 A2 (P21, P22, ..., P2n) 2009-1-1 18:00:00 ... * · * • . Table 3. Update image feature table The above rules ensure that the updated image ID and the last in the image feature table are recorded. The modified time of the record and its updated characteristics値. S 2 3 0 : Associate the extracted updated image feature with the search engine. Specifically, the feature extracted for updating the image may be indexed in the search engine. In fact, the purpose of this step is to build the updated image into the search engine. The BUILD process includes indexing the images to be searched, putting them into memory, and so on. In particular, the feature of each image creates an array on the index. This is because the image feature 値 includes a vector that contains a number of 値 and therefore needs to be stored in an array. For the distributed search engine, the method embodiment may further include: -16- 201113724 s 240: distributing the extracted updated image feature 値 to the distributed search engine. It can be seen from the above embodiment that the updated image is acquired, the feature of the updated image is extracted and stored, and the extracted updated image feature is associated with the search engine, since the entire process only involves the updated image, and The image that has not been updated is processed, and is not processed for the image in the entire image database. Therefore, compared with the prior art, the time required to extract the feature of the updated image can be greatly reduced, and Reduces the time required to correlate features to the search engine, which in turn greatly reduces the time required for the entire update process. The following describes a server embodiment for updating an image in an image database in the present application, and FIG. 3 is a block diagram of the server embodiment, as shown in the figure, including: an obtaining unit 31, configured to acquire a map in which an update occurs. The extraction unit 32 is configured to extract the feature of the updated image and store it; the association unit 3 3 is configured to associate the extracted updated image feature with the search engine. Preferably, in the server, the obtaining unit 31 can obtain an updated image by using the updated image ID in the image update table. The image update table may be stored in a storage unit in the server, and may be stored in a storage unit on other devices than the server, which is not limited herein. Preferably, in the server, the acquiring unit 31 may specifically include a trigger 311' recording unit 312, as shown in FIG. 4, wherein -17-201113724 trigger 311' is used to detect the image database. Whether the image in the image changes; the recording unit 312 is configured to record the ID of the image in which the update occurs when the update command accompanying the update of the image in the image database is detected by the trigger. It should be noted that the composition and connection relationship shown in Fig. 4 above are only an example of the server of the present application. In addition, the database and triggers can be located on top of other devices outside the server, depending on which device is responsible for the image library function. Preferably, in the server, as shown in FIG. 5, the extracting unit 32 may include: a comparing unit 321 configured to compare the image update table with the same image ID as the image feature table by a predetermined period. The first recording unit 322 is configured to extract the feature 该 of the image corresponding to the image ID and record it in the image feature table, and record the current time as a modification when the modification time in the image feature table does not exist. time. Preferably, in the server, as shown in FIG. 6, the extracting unit 32 includes: a comparing unit 321 configured to compare, at a predetermined period, a modification time corresponding to the same image ID of the image feature table; The second recording unit 3 23, if the modification time in the image feature table is earlier than the modification time in the image update table, extract the feature 値 of the image corresponding to the image ID and update it in the image feature table, and update The current time is used as the modification time. -18- 201113724 Preferably, in the server, as shown in FIG. 7, for the distributed search engine, the server further includes: a distribution unit 34 for distributing the extracted updated image features to the distributed Search engine. For the convenience of description, the above devices are described as being separately divided into various units by function. Of course, the functions of the various units can be implemented in the same or multiple software and/or hardware in the practice of the invention. The following is a system for updating an image in an image database of the present application. The system embodiment can be as shown in FIG. 8 and includes a server 8 1 and a search engine 82. The server 81 can be the above. The server of any of Figures 3 to 7; the search engine 82' is for association with the feature of the updated image extracted by the server 81. The following describes another system for updating an image in an image database of the present application, which may be as shown in FIG. 9 and includes a server 9 1 and a user terminal 93, and a search engine 92, wherein the user terminal 93 is used to pass The Internet uploads an image to the server; the server 91 may be any of the servers in FIG. 3 to FIG. 7 above, and also receives and stores the image uploaded by the user terminal 93; the search engine 92 is configured to The feature 値 of the updated image extracted by the server 91 is associated. The following describes another system for updating an image in an image database of the present application, which may be as shown in FIG. 10, including a server 〇i, an internet image capture subsystem 103, and a search engine 1〇2. Among them, -19- 201113724 Internet image capture subsystem 103, used to capture images on the Internet and transmit to the server 1〇1: As mentioned above, the Internet image capture subsystem 103, which may be an image in the Internet that is crawled using a crawler system in search technology. The server 1 〇1 may be any of the above-mentioned servers of FIG. 3 to FIG. 7, and also receives and stores an image transmitted by the internet image capture subsystem 103, and the search engine 102 is configured to The feature 値 of the updated image extracted by the server 101 is associated. As mentioned earlier, the Internet Image Capture Subsystem 1 〇 3 can be an image of the Internet captured by the crawler system in the search technology. It should be noted that the composition and connection relationship shown in Fig. 10 described above are only an example of the system of the present application. In addition, the Internet image capture subsystem 103 of Figure 10 can also be integrated with the server 101 or the search engine 1〇2. As can be seen from the description of the above embodiments, those skilled in the art can clearly understand that the present application can be implemented by means of a software plus a necessary general hardware platform. Based on such understanding, the technical solution of the present application may be embodied in the form of a software product in essence or in a contribution to the prior art, and the computer software product may be stored in a storage medium such as a ROM/RAM or a disk. , optical discs, etc., including instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to perform the methods described in various embodiments or portions of the embodiments of the present application. -20- 201113724 Each embodiment in the present application is described in a progressive manner, and the same similar parts between the various embodiments can be referred to each other. 'Each embodiment focuses on different embodiments from other embodiments. At the office. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the description of the method embodiment. This application can be used in a variety of general or proprietary computing system environments or configurations. For example: personal computers, server computers, handheld devices or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics devices, network PCs, small computers, large Computer, decentralized computing environment including any of the above systems or devices, and so on. This application can be described in the general context of computer-executable instructions executed by a computer, such as a program module. In general, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The invention may also be practiced in a distributed computing environment where tasks are performed by remote processing devices that are connected through a communication network. In a distributed computing environment, the program modules can be located in local and remote computer storage media, including storage devices. While the invention has been described by the embodiments of the present invention, it will be understood that -21 - 201113724 [Simplified description of the drawings] In order to more clearly explain the embodiments of the present application or the technical solutions in the prior art, the following is a brief description of the drawings used in the embodiments or the description of the prior art. The drawings in the following description are only some of the embodiments described in the present application, and those skilled in the art can obtain other drawings according to the drawings without any inventive labor. 1 is a flow chart of an image search solution in the prior art; FIG. 2 is a flow chart of an embodiment of a method for updating an image of the present application; FIG. 3 is a block diagram of an embodiment of a server according to the present application; Figure 5 is a block diagram of another server embodiment of the present application; Figure 6 is a block diagram of another server embodiment of the present application; Figure 7 is another block diagram of the present application; Figure 8 is a block diagram of another system embodiment of the present application; Figure 9 is a block diagram of another system embodiment of the present application; Figure 10 is another system embodiment of the present application. Block diagram. [Major component symbol description] 31: Acquisition unit 32: Extraction unit 33: Association unit 31 1 : Trigger 3 1 2 : Recording unit-22-201113724 3 2 1 : Comparison unit 3 22: First recording unit 3 2 3 : Second recording unit 3 4 : Distribution unit 8 1 : Server 82 : Search engine 9 1 : Server 92 : Search engine 93 : User terminal 1 〇 1 : Server 102 : Search engine 1 03 : Internet image capture Take subsystem -23-