為了使本技術領域的人員更好地理解本說明書一個或多個實施例中的技術方案，下面將結合本說明書一個或多個實施例中的附圖，對本說明書一個或多個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說明書一部分實施例，而不是全部的實施例。基於本說明書一個或多個實施例，本領域普通技術人員在沒有作出創造性勞動前提下所獲得的所有其他實施例，都應當屬於本公開保護的範圍。
本說明書至少一個實施例提供了一種支付處理方法，該處理方法可以使得用戶在付款時不用手動輸入付款金額，提高用戶的體驗。
圖1示例了該方法的一種應用場景，如圖1所示，商戶11購買了一個支付播報設備12，該支付播報設備12可以在支付過程中顯示商戶的收款碼，還可以播報付款方的付款是否到帳，例如，可以播報“您收到付款8元”。
付款方13可以掃描支付播報設備12上顯示的收款碼，向支付伺服器14發送支付請求，由支付伺服器14處理本次支付，並將支付結果發送至支付播報設備12進行播報，以使得商戶11瞭解付款方13是否真正付款到帳。
為了實現不手動輸入金額也能支付，本說明書至少一個實施例的支付處理方法中，將輸入金額的操作交由商家11來實現，並且，由商家11以語音的方式向支付播報設備12輸入金額。詳細的處理可以參見圖2的流程：
在步驟200中，支付播報設備採集收款方的語音信號。
例如，收款方可以是商家，也可以是個人。
以收款方是商家為例，語音信號可以是商家在輸入付款方待支付的付款金額。例如，該語音信號可以是“18元”。
具體實施中，商家可以購買支付播報設備，可以在支付播報設備的說明書中設置一些關於採集語音信號的說明，比如，規定商家在採集語音信號時的輸入格式為“付款金額+元”、或者“付款金額”、或者“付款金額+塊錢”、或者“付款金額+元+整”等。商家可以按照規定的幾種格式進行輸入。例如，當某個用戶要向商家付款時，商家可以按下支付播報設備上的語音採集按鍵，開始輸入語音信號“18元”。
在步驟202中，支付播報設備對語音信號進行識別分析，得到付款金額。
本步驟中，支付播報設備可以對採集的語音信號進行語音識別，識別出其中的付款金額。例如，以上面的例子“18元”為例，在開始本步驟的識別分析之前，支付播報設備可以對語音信號進行預處理，將“18元”中的“元”識別出並去掉，對“18”進行識別分析得到具體的金額數字。
如上所述，當預先規定了商家在採集語音信號時的輸入格式時，還可以在支付播報設備中儲存對這些格式的語音信號進行預處理去掉無關信號的功能模組，該功能模組可以是由軟體程式或者硬體實現。所述的無關信號例如可以包括上述的“元”、“塊錢”、“整”等，留下“18”這些實際數字來進行識別分析。
當然，如上只是一種示例，也可以對步驟200採集的語音信號整體進行分析，比如，對“18元整”開始進行識別分析。其中，支付播報設備的識別分析的過程後續將舉例詳述。假設本步驟識別到付款金額是18。
在步驟204中，支付播報設備將所述付款金額和收款方的帳號標識發送至支付伺服器。
本步驟中，支付播報設備可以將識別到的付款金額“18”以及商家的帳號標識發送至支付伺服器。商家的帳號標識可以是收款方帳號，並且支付播報設備獲取收款方帳號的方式可以有多種，例如，支付播報設備中可以預儲存收款方帳號。或者，還可以透過物聯網設備連接的方式，由商家掃描支付播報設備上的設備標識後向支付伺服器發送關聯請求，將支付播報設備的設備標識和商家的收款方帳號進行關聯，這種方式下，支付播報設備在本步驟中上報給支付伺服器的“收款方的帳號標識”可以是支付播報設備的設備標識，支付伺服器根據設備標識可以找到對應關聯的商家的收款方帳號，因此可以將支付播報設備的設備標識相當於一種“收款方的帳號標識”。
此外，支付播報設備在向支付伺服器發送上述資訊時，可以將付款金額、帳號標識等進行加密傳輸，以保證資訊傳輸的安全性。支付播報設備還可以向支付伺服器發送數位簽章，以提高本次支付校驗的安全性。
在步驟206中，支付伺服器根據付款金額和帳號標識，生成對應的收款碼資訊，並儲存所述收款碼資訊、付款金額和帳號標識三者的對應關係。
例如，支付伺服器在接收到收款方的帳號標識和付款金額後，可以據此生成一個Hash串，其中，如果支付播報設備上傳的帳號標識是支付播報設備的設備標識時，支付伺服器還可以根據設備標識找到用於收款的收款方帳號，並根據收款方帳號和付款金額生成Hash串。或者，支付伺服器也可以根據支付播報設備的設備標識和付款金額生成對應的Hash串，只要後續步驟能夠根據返回的收款碼資訊找到商家用於收款的收款方帳號即可。
當該Hash串較長時，支付伺服器還可以將Hash串轉換成短連結，該短連結用於發送至支付播報設備轉換成收款碼。該短連結可以稱為收款碼資訊。在其他的例子中，收款碼資訊也可以不是短連結，支付伺服器發送至支付播報設備的用於顯示收款碼的資訊就可以稱為收款碼資訊。
此外，支付伺服器還可以儲存收款碼資訊、付款金額和帳號標識的對應關係。同樣的，該對應關係的儲存，收款碼資訊可以是上述的短連結，或者也可以是Hash串，只要能據此找到對應的付款金額和帳號標識即可。而所述的帳號標識可以是真正的商家收款帳號，或者也可以是支付播報設備的設備標識等其他資訊，只要能夠根據帳號標識得到商家用於收款的收款方帳號。
需要說明的是，該收款碼資訊是動態變化的，在每次支付時，當商家向支付播報設備輸入的付款金額不同時，最終得到的收款碼資訊也將不同。
在步驟208中，支付伺服器將收款碼資訊發送至支付播報設備進行顯示。
例如，以收款碼資訊是短連結為例，支付伺服器可以將短連結發送至支付播報設備。
在步驟210中，支付播報設備根據收款碼資訊顯示對應的收款碼。
本步驟中，支付播報設備可以將短連結轉換成收款碼，顯示在支付播報設備的顯示幕上，以供付款方掃描付款。例如，該收款碼可以是二維碼或者其他類型的碼。付款方掃描所述收款碼後，可以得到述收款碼資訊，並將收款碼資訊攜帶在支付請求中，向支付伺服器發送支付請求，用於請求付款。
在步驟212中，支付伺服器接收到支付請求，該請求中攜帶收款碼資訊。
本步驟中，支付伺服器接收到付款方發送的支付請求，該支付請求可以是付款方掃描支付播報設備上的收款碼後發送。此外，該支付請求中還可以攜帶付款方的帳號標識，以使得支付伺服器根據付款方的帳號標識得到付款方帳號。
在步驟214中，支付伺服器根據收款碼資訊查找對應關係，得到付款金額和收款方的帳號標識。
本步驟中，支付伺服器可以查找步驟206中儲存的對應關係，得到收款碼資訊對應的付款金額和收款方的帳號標識，並根據帳號標識得到收款方帳號。
此外，支付伺服器在識別到付款金額和收款方的帳號標識後，可以將收款方的商戶名稱和付款金額傳輸至付款方的手機端，在手機端的支付介面進行顯示。這樣付款方用戶就可以確認下付款金額和商戶名稱是否正確，並在確認無誤後可以輸入支付密碼進行付款。
在步驟216中，支付伺服器進行支付處理。
例如，當付款方用戶在手機端確認支付後，支付伺服器可以由付款方帳號向收款方帳號進行轉帳，轉帳金額是所述的付款金額。
當支付處理成功後，支付伺服器可以將儲存的收款碼資訊、付款金額和帳號標識三者的對應關係進行失效處理，比如，可以將該對應關係由資料庫刪除。
在步驟218中，支付伺服器將支付處理結果發送至支付播報設備，以在支付播報設備進行語音播報。
在步驟220中，支付播報設備語音播報本次支付的支付處理結果。例如，支付播報設備可以透過播放喇叭播報“收到付款18元”。透過在支付完成後，將支付處理結果推送到支付播報設備，設備進行播報，從而保證不會出現逃單而未發現的現象，解決了線下逃單的問題。
本說明書至少一個實施例的支付處理方法，透過由收款方向支付播報設備語音輸入付款金額，使得收款方不用手動輸入金額，而且付款方掃描支付播報設備上的收款碼就可以進行付款，也不用手動輸入金額，實現了在支付過程中無需商家或用戶手動輸入金額，提高了支付的便利性。同時，該方法中的收款碼是隨著每次付款金額的不同而動態變化的，是一種動態收款碼，可以在很大程度上避免對收款碼的替換，提高了支付安全性。
在圖2所示的例子中，語音信號的採集和付款金額的識別，可以是在支付播報設備執行。在其他的例子中，圖2所示流程中的各個處理步驟的執行者也可以進行變動。例如，支付播報設備可以負責採集收款方的語音信號，然後可以將語音信號發送至雲端的支付伺服器進行識別分析得到付款金額。支付播報設備可以將收款方的帳號標識發送至支付伺服器，待支付伺服器識別到付款金額後，可以根據付款金額和帳號標識生成收款碼資訊。
本說明書的至少一個實施例中，支付播報設備採集收款方的語音信號並進行識別分析，可以採用多種語音識別方法獲得語音信號中的付款金額。如下示例一種語音識別的處理過程，如圖3所示，可以包括如下處理，但是實際實施中並不局限於如下各步驟的處理順序：
在步驟300中，支付播報設備的麥克風採集收款方的語音信號。
例如，支付播報設備中用於採集收款方語音的麥克風，可以是單麥克風或者麥克風陣列。該麥克風陣列可以是二維麥克風陣列或者線性麥克風等。麥克風陣列是由一定數目的聲學感測器組成，用來對聲場的空間特性進行採樣並處理的系統。麥克風陣列中可以包括多個陣元。
在步驟302中，支付播報設備識別所述語音信號。
以麥克風陣列採集為例，本步驟中，可以根據麥克風陣列中的各個陣元採集的語音得到整個陣列輸出的語音信號。
首先，可以透過廣義互相關方法確定各個陣元的時延。例如，可以尋找廣義互相關函數中的峰值點，透過峰值點的位置計算出各個陣元的時延。
在下面的公式(1)中，n是陣元，x(n)是第n個陣元採集到的語音，m是第n個陣元對應的時延。當將公式(1)的值最大時，得到m。對於不同的陣元，可以得到分別對應該陣元的時延m。
其中，在上述的公式(1)中，X(n)表示參照陣元的語音，y(n)表示要計算偏移量的陣元的語音。y(n-m)表示在y(n)的基礎上進行了偏移。
接著，當得到各個陣元分別對應的時延m後，可以按照公式(2)計算整個麥克風陣列識別輸出的語音信號。假如商家輸入的語音是“18”，那麼可以將這個語音信號識別出來，不過還未識別到具體的金額數字。
在上面的公式(2)中，y(n)是最終識別到的語音信號，k是陣元數，該陣列一共有k個陣元，mk
是第k個陣元對應的時延。透過確定陣列中的各個陣元的時延，並將各個陣元採集的語音信號進行求和平均，得到整個麥克風陣列的語音信號，可以降低雜訊的影響，使得在嘈雜環境下也能提取到商家輸入的語音。此外，在支付播報設備中設置麥克風陣列的成本也較低。
在步驟304中，支付播報設備對語音信號進行語音識別，得到音素序列。
本步驟中，可以對步驟302中得到語音信號進行識別得到音素序列。其中，音素是語音中的最小的單位，依據音節裡的發音動作來分析，一個動作構成一個音素。音素分為母音、輔音兩大類。如漢語音節ā(啊)只有一個音素，ài(愛)有兩個音素，dāi(呆)有三個音素等。音素序列中包括多個音素，例如，dāi(呆)有三個音素，d、a、i可以稱為一個音素序列。
在步驟306中，對語音信號計算語音特徵。
本步驟和步驟308這兩個步驟，可以是混淆矩陣的生成。混淆矩陣可以是離線計算生成，可以是根據對標注語音資料集進行音素識別和音素聚類得到。比如，可以獲取一些現有的標注資料集，該標注資料集可以是之前已經採集過的語音。例如，該標注資料集中可以包括正常的數字金額的語音以及容易被誤識別到的語音，比如，6(liu)有出現過被誤識別為“牛”(niu)。對標注資料集中的語音進行音素識別，對計算出錯的音素，統計混淆率。可以對步驟302中得到的語音信號進行分幀、加窗、傅立葉變換等處理，計算語音特徵，例如，可以得到梅爾倒譜系數(Mel-scale Frequency Cepstral Coefficients，簡稱MFCC)特徵或者線性預測倒譜系數(Linear Prediction Cepstrum Coefficient, LPCC)特徵。
在步驟308中，根據語音特徵進行音素聚類得到音素聚類資訊。
在得到語音特徵的基礎上，可以根據上述的語音特徵進行音素的聚類，得到易混淆音素的聚類。如“l”和“n”為易混淆音素。本步驟可以得到混淆矩陣，混淆矩陣中的參數表示第i個音素被混淆為第j個音素的機率。比如，音素“l”容易被混淆成“n”，且有2%的機率被混淆為“n”，那麼混淆矩陣中的其中一個參數表示音素“l”被混淆為“n”的機率是0.02。
在步驟310中，根據音素聚類資訊對音素序列擴展得到多個待選音素序列。
本步驟中，可以對步驟304得到的音素序列，透過上述的音素聚類資訊(如，步驟308得到的混淆矩陣)進行擴展，得到多個待選音素陣列。由於步驟304得到的音素序列可能存在誤差，比如，一些發音比較相似的音素之間出現混淆，所以根據混淆矩陣，得到了多個可能的音素序列，以盡可能的使得識別到的音素序列更加準確。由於支付金額的文本相對較少，透過上述的根據混淆矩陣來擴展音素序列的方式，可以將相近的音素歸類到多個待選音素陣列中，提升對付款金額的識別準確性。
在步驟312中，根據數字文法先驗知識，由所述多個待選音素序列中得到目標音素序列。本步驟中，所述的數字文法先驗知識指的是在數字金額的識別中最有可能出現的音素是什麼，或者說，根據數字金額的發音來看，哪些音素具有較大機率的是數字金額的音素。
舉例來說，數字金額的數量是有限的，1、2、3、4等，以6為例，該數字6的發音liu對應的音素序列可以是“l、i、u”，而“l”與“n”這兩個音素在發音時容易混淆，屬於易混淆音素，可能會在音素識別時識別為音素序列“n、i、u”。在進行音素序列的擴展時，假設得到了兩個待選音素序列“n、i、u”和“l、i、u”。根據數字文法先驗知識，音素序列“n、i、u”組合出來的發音niu基本不可能是一個數字金額，所以可以選擇“l、i、u”作為最終得到的目標音素序列。這個目標音素序列是機率最高組成數字金額的序列。
透過上面的例子可以看到，透過音素聚類的混淆矩陣，可以對語音識別到的音素序列進行擴展，使得發音鄰近的音素也被歸類到可能的數字金額序列中，再根據數字文法先驗知識從中選擇一個機率最高的音素序列，從而提高了對文本相對較少的付款金額識別的準確性。
在步驟314中，根據目標音素序列得到付款金額。
例如，假設識別到的目標音素序列是“l、i、u”，可以得到付款金額是6。
透過基於麥克風陣列進行語音信號採集，並進行音素的識別和糾錯，得到最終的識別付款金額，使得付款金額的識別更加準確。
為了實現上述方法，本說明書至少一個實施例提供了一種支付播報設備，如圖4所示，該設備可以包括：語音識別模組41、資訊收發模組42和顯示處理模組43。
語音識別模組41，用於採集收款方的語音信號，所述語音信號是收款方在輸入付款方待支付的付款金額；對語音信號進行識別分析得到所述付款金額；
資訊收發模組42，用於將所述付款金額和所述收款方的帳號標識，發送至支付伺服器；接收所述支付伺服器返回的收款碼資訊，所述收款碼資訊是支付伺服器根據付款金額和帳號標識對應生成；
顯示處理模組43，用於在顯示幕上顯示所述收款碼資訊對應的收款碼，以使得付款方掃描所述收款碼向支付伺服器發送付款請求。
在一個例子中，語音識別模組41，包括：麥克風和識別分析子模組；
所述麥克風，用於採集收款方的語音信號；
所述識別分析子模組，用於識別透過麥克風採集的語音信號；對所述語音信號進行語音識別，得到音素序列；根據音素聚類資訊對所述音素序列擴展，得到多個待選音素序列；所述音素聚類資訊是根據對標注語音資料集進行音素識別和音素聚類得到；根據數字文法先驗知識，由所述多個待選音素序列中得到目標音素序列；根據所述目標音素序列得到所述付款金額。
為了實現上述方法，本說明書至少一個實施例提供了一種支付伺服器，如圖5所示，該伺服器可以包括：資訊接收模組51、碼處理模組52和資訊發送模組53。
資訊接收模組51，用於接收支付播報設備發送的付款金額和收款方的帳號標識，所述付款金額是由支付播報設備採集收款方的語音信號並識別分析得到；
碼處理模組52，用於根據付款金額和帳號標識，生成對應的收款碼資訊；
資訊發送模組53，用於將所述收款碼資訊發送至支付播報設備，以顯示在所述支付播報設備。
在一個例子中，資訊接收模組51，還用於將所述收款碼資訊、付款金額和帳號標識的對應關係，進行儲存；當收到支付請求時，根據支付請求中攜帶的收款碼資訊，查找所述對應關係得到對應的付款金額和帳號標識；並在支付處理成功後，將儲存的所述對應關係，做失效處理。
上述實施例闡明的裝置或模組，具體可以由電腦晶片或實體實現，或者由具有某種功能的產品來實現。一種典型的實現設備為電腦，電腦的具體形式可以是個人電腦、膝上型電腦、行動電話、相機電話、智慧型電話、個人數位助理、媒體播放機、導航設備、電子郵件收發設備、遊戲控制台、平板電腦、可穿戴設備或者這些設備中的任意幾種設備的組合。
為了描述的方便，描述以上裝置時以功能分為各種模組分別描述。當然，在實施本說明書一個或多個實施例時可以把各模組的功能在同一個或多個軟體和/或硬體中實現。
上述圖中所示流程中的各個步驟，其執行順序不限制於流程圖中的順序。此外，各個步驟的描述，可以實現為軟體、硬體或者其結合的形式，例如，本領域技術人員可以將其實現為軟體代碼的形式，可以為能夠實現所述步驟對應的邏輯功能的電腦可執行指令。當其以軟體的方式實現時，所述的可執行指令可以儲存在記憶體中，並被設備中的處理器執行。
例如，對應於上述方法，本說明書一個或多個實施例同時提供一種支付播報設備。該設備可以包括處理器、記憶體、以及儲存在記憶體上並可在處理器上運行的電腦指令，所述處理器透過執行所述指令，用於實現如下步驟：
採集收款方的語音信號，所述語音信號是收款方在輸入付款方待支付的付款金額；
對所述語音信號進行識別分析，得到所述付款金額；
將所述付款金額和所述收款方的帳號標識，發送至支付伺服器；
接收所述支付伺服器返回的收款碼資訊，所述收款碼資訊是支付伺服器根據付款金額和帳號標識對應生成；
根據所述收款碼資訊顯示對應的收款碼，以使得付款方掃描所述收款碼向支付伺服器發送支付請求。
例如，對應於上述方法，本說明書一個或多個實施例同時提供一種支付伺服器。該伺服器可以包括處理器、記憶體、以及儲存在記憶體上並可在處理器上運行的電腦指令，所述處理器透過執行所述指令，用於實現如下步驟：
接收支付播報設備發送的付款金額和收款方的帳號標識，所述付款金額是由支付播報設備採集收款方的語音信號並識別分析得到；
根據所述付款金額和帳號標識，生成對應的收款碼資訊；
將所述收款碼資訊發送至支付播報設備，以顯示在所述支付播報設備。
還需要說明的是，術語“包括”、“包含”或者其任何其他變體意在涵蓋非排他性的包含，從而使得包括一系列要素的過程、方法、商品或者設備不僅包括那些要素，而且還包括沒有明確列出的其他要素，或者是還包括為這種過程、方法、商品或者設備所固有的要素。在沒有更多限制的情況下，由語句“包括一個……”限定的要素，並不排除在包括所述要素的過程、方法、商品或者設備中還存在另外的相同要素。
本領域技術人員應明白，本說明書一個或多個實施例可提供為方法、系統或電腦程式產品。因此，本說明書一個或多個實施例可採用完全硬體實施例、完全軟體實施例或結合軟體和硬體方面的實施例的形式。而且，本說明書一個或多個實施例可採用在一個或多個其中包含有電腦可用程式碼的電腦可用儲存媒體(包括但不限於磁碟記憶體、CD-ROM、光學記憶體等)上實施的電腦程式產品的形式。
本說明書一個或多個實施例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。也可以在分散式運算環境中實踐本說明書一個或多個實施例，在這些分散式運算環境中，由透過通信網路而被連接的遠端處理設備來執行任務。在分散式運算環境中，程式模組可以位於包括存放裝置在內的本地和遠端電腦儲存媒體中。
本說明書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，對於資料處理設備實施例而言，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。
上述對本說明書特定實施例進行了描述。其它實施例在所附申請專利範圍的範圍內。在一些情況下，在申請專利範圍中記載的動作或步驟可以按照不同於實施例中的順序來執行並且仍然可以實現期望的結果。另外，在附圖中描繪的過程不一定要求示出的特定順序或者連續順序才能實現期望的結果。在某些實施方式中，多工處理和並行處理也是可以的或者可能是有利的。
以上所述僅為本說明書一個或多個實施例的較佳實施例而已，並不用以限制本說明書一個或多個實施例，凡在本說明書一個或多個實施例的精神和原則之內，所做的任何修改、等同替換、改進等，均應包含在本說明書一個或多個實施例保護的範圍之內。In order to enable those skilled in the art to better understand the technical solutions in one or more embodiments of this specification, the following will be combined with the drawings in one or more embodiments of this specification. The technical solution is described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of this specification, but not all the embodiments. Based on one or more embodiments of this specification, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present disclosure. At least one embodiment of the present specification provides a payment processing method, which can enable a user to avoid entering a payment amount manually when making a payment, thereby improving the user's experience. FIG. 1 illustrates an application scenario of the method. As shown in FIG. 1, the merchant 11 purchases a payment broadcast device 12, which can display the merchant’s collection code during the payment process, and can also broadcast the payer’s payment code. Whether the payment is received, for example, you can broadcast "You received a payment of 8 yuan." The payer 13 can scan the payment code displayed on the payment reporting device 12, send a payment request to the payment server 14, the payment server 14 processes the payment, and sends the payment result to the payment reporting device 12 for broadcasting, so that The merchant 11 knows whether the payer 13 has actually paid to the account. In order to realize payment without manually inputting the amount, in the payment processing method of at least one embodiment of the present specification, the operation of inputting the amount is handed over to the merchant 11, and the merchant 11 enters the amount into the payment broadcasting device 12 by voice . For the detailed processing, please refer to the flow of FIG. 2: In step 200, the payment broadcasting device collects the payee's voice signal. For example, the payee can be a merchant or an individual. Taking the example that the payee is a merchant, the voice signal may be that the merchant enters the payment amount to be paid by the payer. For example, the voice signal may be "18 yuan". In the specific implementation, the merchant can purchase the payment broadcast device, and can set some instructions on collecting the voice signal in the specification of the payment broadcast device, for example, stipulating that the input format of the merchant when collecting the voice signal is "payment amount + yuan", or " Payment Amount", or "Payment Amount+Cash", or "Payment Amount+Yuan+Integer", etc. Merchants can enter according to the prescribed formats. For example, when a user wants to pay a merchant, the merchant can press the voice collection button on the payment announcement device and start to input the voice signal "18 yuan". In step 202, the payment broadcasting device performs recognition analysis on the voice signal to obtain the payment amount. In this step, the payment broadcast device can perform voice recognition on the collected voice signal, and recognize the payment amount therein. For example, taking the above example "18 yuan" as an example, before starting the recognition analysis in this step, the payment broadcasting device can preprocess the voice signal, identify and remove the "yuan" from the "18 yuan", and correct the " 18” Perform identification analysis to get specific amount figures. As mentioned above, when the input format of the merchant when collecting voice signals is predetermined, a functional module that preprocesses voice signals in these formats to remove irrelevant signals may also be stored in the payment broadcast device, and the functional module may be Realized by software program or hardware. The irrelevant signal may include, for example, the above-mentioned "yuan", "dollar", "integer", etc., leaving the actual numbers "18" for identification and analysis. Of course, the above is just an example, and the entire voice signal collected in step 200 may also be analyzed, for example, the recognition analysis of "18 yuan integer" is started. Among them, the process of identification and analysis of the payment broadcast equipment will be detailed later by way of example. Assume that the payment amount identified in this step is 18. In step 204, the payment broadcast device sends the payment amount and the account identification of the payee to the payment server. In this step, the payment broadcast device may send the identified payment amount "18" and the merchant's account identification to the payment server. The account ID of the merchant may be the payee account number, and there may be multiple ways for the payment reporting device to obtain the payee account number. For example, the pay reporting device may pre-store the payee account number. Or, it can also be connected through an Internet of Things device, and the merchant scans the device identification on the payment broadcast device and sends an association request to the payment server to associate the device identification of the payment broadcast device with the merchant's payee account. In this mode, the “billee account ID” reported by the payment broadcast device to the payment server in this step may be the device identification of the payment broadcast device, and the payment server may find the payee of the associated merchant according to the device identification Account number, so the device identification of the payment broadcasting device can be equivalent to a kind of "account identification of the payee". In addition, when sending the above information to the payment server, the payment broadcasting device can encrypt the payment amount, account identification, etc. to ensure the security of the information transmission. The payment broadcast device can also send a digital signature to the payment server to improve the security of this payment verification. In step 206, the payment server generates corresponding payment code information according to the payment amount and account identification, and stores the correspondence between the collection code information, the payment amount, and the account identification. For example, after receiving the account identification and payment amount of the payee, the payment server may generate a hash string accordingly. If the account identification uploaded by the payment reporting device is the device identification of the payment reporting device, the payment server also You can find the beneficiary account number used for receiving money according to the device ID, and generate a hash string based on the beneficiary account number and payment amount. Alternatively, the payment server can also generate a corresponding Hash string based on the device identification and payment amount of the payment broadcast device, as long as the subsequent steps can find the payee account used by the merchant to receive the payment based on the returned collection code information. When the hash string is long, the payment server can also convert the hash string into a short link, which is used to send to the payment broadcast device and convert it into a collection code. This short link can be called the collection code information. In other examples, the payment code information may not be a short link. The information that the payment server sends to the payment broadcasting device to display the payment code may be called the payment code information. In addition, the payment server can also store the correspondence between the collection code information, payment amount and account ID. Similarly, the storage of the corresponding relationship, the collection code information can be the above short link, or it can also be a hash string, as long as the corresponding payment amount and account identification can be found based on this. The account identification can be a real merchant collection account, or it can also be other information such as the device identification of the payment broadcast device, as long as the payee account used by the merchant for collection can be obtained according to the account identification. It should be noted that the payment code information changes dynamically. In each payment, when the payment amount input by the merchant to the payment broadcasting device is different, the received payment code information will also be different. In step 208, the payment server sends the payment code information to the payment broadcast device for display. For example, taking the receipt code information as a short link as an example, the payment server can send the short link to the payment broadcast device. In step 210, the payment broadcasting device displays the corresponding payment code according to the payment code information. In this step, the payment broadcast device can convert the short link into a collection code and display it on the display screen of the payment broadcast device for the payer to scan the payment. For example, the payment code may be a two-dimensional code or other types of codes. After scanning the payment code, the payer can obtain the payment code information, and carry the payment code information in the payment request, and send the payment request to the payment server for requesting payment. In step 212, the payment server receives the payment request, and the request carries the payment code information. In this step, the payment server receives the payment request sent by the payer. The payment request may be sent by the payer after scanning the payment code on the payment broadcast device. In addition, the payment request may also carry the account ID of the payer, so that the payment server obtains the account number of the payer according to the account ID of the payer. In step 214, the payment server looks up the corresponding relationship according to the payment code information to obtain the payment amount and the account ID of the payee. In this step, the payment server may look up the correspondence stored in step 206 to obtain the payment amount corresponding to the payment code information and the account ID of the payee, and obtain the account number of the payee according to the account ID. In addition, after the payment server recognizes the payment amount and the payee's account ID, it can transmit the payee's merchant name and payment amount to the payer's mobile phone and display it on the mobile phone's payment interface. In this way, the user of the payer can confirm whether the payment amount and the merchant name are correct, and after confirming the correctness, they can enter the payment password to make the payment. In step 216, the payment server performs payment processing. For example, after the payer user confirms the payment on the mobile phone, the payment server may transfer the money from the payer account to the payee account, and the transfer amount is the payment amount. When the payment processing is successful, the payment server can invalidate the correspondence between the stored payment code information, the payment amount, and the account ID. For example, the correspondence can be deleted from the database. In step 218, the payment server sends the payment processing result to the payment broadcast device to perform voice broadcast on the payment broadcast device. In step 220, the payment broadcasting device voice broadcasts the payment processing result of this payment. For example, the payment broadcast device can broadcast the "18 yuan received payment" through the broadcast speaker. After the payment is completed, the payment processing result is pushed to the payment broadcast device, and the device broadcasts, thereby ensuring that there is no phenomenon of escaping the order and not discovering it, and solving the problem of offline escaping the order. In the payment processing method of at least one embodiment of the present specification, the payment amount is input from the payment direction to the payment broadcasting device by the receiving party, so that the payee does not need to manually enter the amount, and the payment party scans the payment code on the payment broadcasting device to make the payment. There is no need to manually input the amount, which realizes that there is no need for the merchant or the user to manually input the amount during the payment process, which improves the convenience of payment. At the same time, the payment code in this method changes dynamically with the amount of each payment. It is a dynamic payment code, which can largely avoid the replacement of the payment code and improve the payment security. In the example shown in FIG. 2, the collection of voice signals and the identification of payment amounts may be performed at the payment broadcast device. In other examples, the executor of each processing step in the flow shown in FIG. 2 may also be changed. For example, the payment broadcast device may be responsible for collecting the payee's voice signal, and then may send the voice signal to the cloud's payment server for recognition and analysis to obtain the payment amount. The payment broadcast device can send the account identification of the payee to the payment server. After the payment server recognizes the payment amount, it can generate the payment code information according to the payment amount and the account identification. In at least one embodiment of this specification, the payment broadcast device collects the payee's voice signal and performs recognition analysis, and various payment methods can be used to obtain the payment amount in the voice signal. The following is an example of a speech recognition processing procedure, as shown in FIG. 3, which may include the following processing, but the actual implementation is not limited to the processing sequence of the following steps: In step 300, the microphone of the payment broadcasting device collects the payee’s voice signal. For example, the microphone used to collect the payee's voice in the payment broadcasting device may be a single microphone or a microphone array. The microphone array may be a two-dimensional microphone array or a linear microphone. The microphone array is composed of a certain number of acoustic sensors, used to sample and process the spatial characteristics of the sound field. Multiple array elements may be included in the microphone array. In step 302, the payment announcement device recognizes the voice signal. Taking the microphone array as an example, in this step, the voice signal output by the entire array can be obtained according to the voice collected by each element in the microphone array. First, the delay of each array element can be determined by the generalized cross-correlation method. For example, you can find the peak point in the generalized cross-correlation function, and calculate the time delay of each array element through the position of the peak point. In the following formula (1), n is the array element, x(n) is the speech collected by the nth array element, and m is the delay corresponding to the nth array element. When the value of formula (1) is maximized, m is obtained. For different array elements, the delay m corresponding to each array element can be obtained. In the above formula (1), X(n) represents the voice of the reference element, and y(n) represents the voice of the element whose offset is to be calculated. y(nm) indicates that an offset has been made on the basis of y(n). Next, when the time delays m corresponding to the respective array elements are obtained, the voice signal recognized by the entire microphone array can be calculated according to formula (2). If the voice input by the merchant is "18", then this voice signal can be recognized, but the specific amount number has not yet been recognized. In the above formula (2), y(n) is the finally recognized speech signal, k is the number of array elements, the array has a total of k array elements, and m k is the delay corresponding to the kth array element. By determining the time delay of each element in the array, and summing and averaging the voice signals collected by each element, the voice signal of the entire microphone array can be obtained, which can reduce the influence of noise and enable extraction in noisy environments. The voice entered by the business. In addition, the cost of setting up the microphone array in the payment broadcast equipment is also low. In step 304, the payment broadcasting device performs speech recognition on the speech signal to obtain a phoneme sequence. In this step, the speech signal obtained in step 302 may be identified to obtain a phoneme sequence. Among them, the phoneme is the smallest unit in speech, which is analyzed according to the pronunciation actions in the syllable, and one action constitutes one phoneme. Phonemes are divided into vowels and consonants. For example, Chinese syllable ā (ah) has only one phoneme, ài (love) has two phonemes, and dāi (stay) has three phonemes. The phoneme sequence includes multiple phonemes. For example, dāi (stay) has three phonemes. d, a, and i can be called a phoneme sequence. In step 306, speech features are calculated on the speech signal. The two steps of this step and step 308 may be the generation of a confusion matrix. The confusion matrix can be generated offline, and can be obtained by phoneme recognition and phoneme clustering on the labeled speech data set. For example, some existing annotation data sets may be obtained, and the annotation data sets may be voices that have been previously collected. For example, the labeling data set may include a normal digital amount of voice and a voice that is easily misrecognized. For example, 6 (liu) has been misrecognized as "niu". Phoneme recognition is performed on the voices in the marked data set, and the phonetic errors in the calculation are calculated, and the confusion rate is counted. The speech signal obtained in step 302 may be subjected to processing such as framing, windowing, and Fourier transform to calculate speech characteristics. For example, Mel-scale Frequency Cepstral Coefficients (MFCC) characteristics or linear prediction inversion may be obtained Spectral coefficient (Linear Prediction Cepstrum Coefficient, LPCC) feature. In step 308, phoneme clustering is performed according to speech features to obtain phoneme clustering information. On the basis of obtaining speech features, phoneme clustering can be performed according to the above speech features to obtain clusters of phonemes that are easily confused. For example, "l" and "n" are confusing phonemes. This step can get the confusion matrix, the parameters in the confusion matrix Indicates the probability that the ith phoneme is confused with the jth phoneme. For example, the phoneme "l" is easily confused as "n", and there is a 2% chance of being confused as "n", then one of the parameters in the confusion matrix The probability that the phoneme "l" is confused as "n" is 0.02. In step 310, the phoneme sequence is expanded according to the phoneme clustering information to obtain a plurality of phoneme sequences to be selected. In this step, the phoneme sequence obtained in step 304 can be expanded through the above phoneme clustering information (eg, the confusion matrix obtained in step 308) to obtain a plurality of phoneme arrays to be selected. Since there may be errors in the phoneme sequence obtained in step 304, for example, some phonemes with similar pronunciations are confused, so according to the confusion matrix, multiple possible phoneme sequences are obtained to make the recognized phoneme sequence more accurate as possible . Since the text of the payment amount is relatively small, by extending the phoneme sequence according to the confusion matrix, similar phonemes can be classified into a plurality of phoneme arrays to be selected, which improves the recognition accuracy of the payment amount. In step 312, the target phoneme sequence is obtained from the plurality of candidate phoneme sequences according to the prior knowledge of digital grammar. In this step, the a priori knowledge of digital grammar refers to what is the most likely phoneme to appear in the recognition of the digital amount, or, according to the pronunciation of the digital amount, which phonemes have a higher probability of being a number Amount of phonemes. For example, the number of digital amounts is limited, 1, 2, 3, 4, etc. Taking 6 as an example, the phoneme sequence corresponding to the pronunciation of the number 6 can be "l, i, u", and "l" The two phonemes that are "n" are easily confused when they are pronounced. They are easy to be confused, and may be recognized as the phoneme sequence "n, i, u" during phoneme recognition. When expanding the phoneme sequence, it is assumed that two candidate phoneme sequences "n, i, u" and "l, i, u" are obtained. According to the prior knowledge of digital grammar, the pronunciation niu of the phoneme sequence "n, i, u" is basically impossible to be a digital amount, so you can choose "l, i, u" as the final target phoneme sequence. This target phoneme sequence is the sequence with the highest probability of forming a digital amount. Through the above example, we can see that through the confusion matrix of phoneme clustering, the phoneme sequence recognized by the speech can be expanded, so that the phonemes adjacent to the pronunciation are also classified into the possible digital amount sequence, and then according to the digital grammar prior The knowledge selects a phoneme sequence with the highest probability, thereby improving the accuracy of identifying the payment amount with relatively small text. In step 314, the payment amount is obtained according to the target phoneme sequence. For example, assuming that the recognized target phoneme sequence is "l, i, u", it can be obtained that the payment amount is 6. Through the collection of voice signals based on the microphone array, and the identification and correction of phonemes, the final recognition payment amount is obtained, which makes the payment amount recognition more accurate. In order to implement the above method, at least one embodiment of this specification provides a payment broadcast device. As shown in FIG. 4, the device may include: a voice recognition module 41, an information transceiver module 42, and a display processing module 43. The voice recognition module 41 is used to collect the payee's voice signal, which is the payee's input of the payment amount to be paid by the payer; the voice signal is identified and analyzed to obtain the payment amount; the information receiving and sending module 42: used to send the payment amount and the account identification of the payee to the payment server; receive the payment code information returned by the payment server, the payment code information is the payment server according to the payment The amount and the account identification are generated correspondingly; the display processing module 43 is used to display the collection code corresponding to the collection code information on the display screen, so that the payer scans the collection code to send a payment request to the payment server. In one example, the voice recognition module 41 includes: a microphone and a recognition and analysis sub-module; the microphone is used to collect the voice signal of the payee; and the recognition and analysis sub-module is used to recognize Speech signal; perform speech recognition on the speech signal to obtain a phoneme sequence; expand the phoneme sequence according to phoneme clustering information to obtain a plurality of phoneme sequences to be selected; the phoneme clustering information is based on the labeled speech data set Phoneme recognition and phoneme clustering are obtained; according to prior knowledge of digital grammar, a target phoneme sequence is obtained from the plurality of candidate phoneme sequences; and the payment amount is obtained according to the target phoneme sequence. In order to implement the above method, at least one embodiment of this specification provides a payment server. As shown in FIG. 5, the server may include: an information receiving module 51, a code processing module 52, and an information sending module 53. The information receiving module 51 is used to receive the payment amount sent by the payment broadcast device and the account identification of the payee. The payment amount is obtained by the payment broadcast device collecting the payee's voice signal and identifying and analyzing it; code processing module 52 , Used to generate corresponding payment code information according to the payment amount and account identification; the information sending module 53 is used to send the payment code information to the payment broadcast device for display on the payment broadcast device. In one example, the information receiving module 51 is also used to store the correspondence between the payment code information, the payment amount and the account identification; when a payment request is received, according to the payment code carried in the payment request Information, search for the corresponding relationship to obtain the corresponding payment amount and account ID; and after the payment processing is successful, the stored corresponding relationship is invalidated. The device or module explained in the above embodiments may be realized by a computer chip or entity, or by a product with a certain function. A typical implementation device is a computer, and the specific form of the computer may be a personal computer, a laptop computer, a mobile phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email sending and receiving device, and a game control Desk, tablet, wearable device, or any combination of these devices. For the convenience of description, when describing the above device, the functions are divided into various modules and described separately. Of course, when implementing one or more embodiments of this specification, the functions of each module may be implemented in one or more software and/or hardware. The execution order of each step in the flow shown in the above figure is not limited to the order in the flowchart. In addition, the description of each step can be implemented in the form of software, hardware, or a combination thereof. For example, those skilled in the art can implement it in the form of software code, which can be a computer capable of implementing the logical functions corresponding to the steps. Execute instructions. When implemented in software, the executable instructions can be stored in memory and executed by the processor in the device. For example, corresponding to the above method, one or more embodiments of this specification simultaneously provide a payment broadcast device. The device may include a processor, a memory, and computer instructions stored on the memory and executable on the processor. The processor executes the instructions to implement the following steps: collect the voice signal of the payee , The voice signal is the payee entering the payment amount to be paid by the payer; the voice signal is identified and analyzed to obtain the payment amount; the payment amount and the account ID of the payee are sent To the payment server; receiving the payment code information returned by the payment server, the payment code information is generated by the payment server according to the payment amount and the account ID; displaying the corresponding payment code according to the payment code information , So that the payer scans the payment code and sends a payment request to the payment server. For example, corresponding to the above method, one or more embodiments of this specification simultaneously provide a payment server. The server may include a processor, a memory, and computer instructions stored on the memory and executable on the processor. The processor executes the instructions to implement the following steps: receiving the payment broadcast device The payment amount and the account identification of the payee. The payment amount is obtained by the payment broadcast device collecting the payee's voice signal and identifying and analyzing; according to the payment amount and account identification, the corresponding payment code information is generated; The payment code information is sent to the payment broadcast device to be displayed on the payment broadcast device. It should also be noted that the terms "include", "include" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device that includes a series of elements includes not only those elements, but also includes Other elements not explicitly listed, or include elements inherent to this process, method, commodity, or equipment. Without more restrictions, the element defined by the sentence "include one..." does not exclude that there are other identical elements in the process, method, commodity, or equipment that includes the element. Those skilled in the art should understand that one or more embodiments of this specification may be provided as a method, system, or computer program product. Therefore, one or more embodiments of this specification may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, one or more embodiments of this specification can be implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) containing computer-usable program code In the form of computer program products. One or more embodiments of this specification may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. One or more embodiments of this specification can also be practiced in a distributed computing environment in which remote processing devices connected through a communication network perform tasks. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices. The embodiments in this specification are described in a progressive manner. The same or similar parts between the embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments. In particular, for the embodiment of the data processing device, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant part can be referred to the description of the method embodiment. The foregoing describes specific embodiments of the present specification. Other embodiments are within the scope of the attached patent application. In some cases, the actions or steps described in the scope of the patent application may be performed in a different order than in the embodiment and still achieve the desired result. In addition, the processes depicted in the drawings do not necessarily require the particular order shown or sequential order to achieve the desired results. In some embodiments, multiplexing and parallel processing are also possible or may be advantageous. The above are only preferred embodiments of one or more embodiments of this specification, and are not intended to limit one or more embodiments of this specification. Anything within the spirit and principle of one or more embodiments of this specification, Any modifications, equivalent replacements, improvements, etc. made should be included within the scope of protection of one or more embodiments of this specification.
