經濟部中央揉率局負工消費合作社印家 4344 96 五、發明説明（1) --- 本發明係有關於-種資料輸人裝置，特別係有關於 一種攝影式智慧閱讀裝置及其操作方法。 按現今之開放社會，媒體資訊日益發達，尤以報紙、 雜諸、書刊為最’大4的新資訊時常給_者造成理解 上的困難，如此，查詢相關字彙的資料或解釋已迫在眉 睫。按以往常見的方法，—般是搬來沉重的字典進行手 工查閱，以達理解的目的。此方法費時費力如遇一篇 資訊材料有多處生僻字彙，則勢必降低閱讀速度，且分 散閱讀者的注意力。 於是，近來出現了許多電腦字彙翻譯軟體，來取代 部份或全部字惫解釋之功能《惟目前所發展出來的字彙 翻譯軟體，在解釋印刷體字彙時，均需將其經由鍵盤或 筆式鍵入的動作來達到輸入的目的，雖省卻人為查閱字 典的步驟，但增加了鍵入字彙的動作’如遇較長的字彙， 速度反不及手工查閲，且鍵入過程中難免發生拼寫錯 誤，使用者常需修改，異常頻繁，極易擾亂閲讀思路， 影響閱讀效果，實為得不償失。 有鑑於此’本發明之目的即在於克服印刷體字彙翻 譯快速性與便利性之矛盾，結合電子技術及程式設計， 提供一攝影式閱讀裝置。此種攝影式閱讀裝置係包含有 一個二維CCD或CMOS光影像感應裝置（CMOS image sensor module) ’其操作步驟及内部控制流程為：（1)首先 使用者將攝影式閱讀與主電腦裝置經由高速傳輸纜線連 接’再將攝影式閱讀裝置之前端透鏡對準待查詢之印刷 3 本紙張尺度適用中國國家標準（CNS ) A4規格（210X297公廣） ^1訂 ^ (請先閲讀背面之注意事項再填寫本頁} 經濟部中央標率局貝工消費合作社印装 4344 96 Λ7 B7____ 五、發明説明（2 ) 體文字或片語，按動採集開關，經由光攝影透鏡擷取該 文字影像的光信號，同時並傳送至攝影式閱讀裝置内的 影像感應裝置（CCD或CMOS影像感應裝置）；（2)在前述 影像感應裝置内部，攝像驅動器（image driver)經由數位 信號處理器控制，以驅動影像感應器將光信號轉換成類 比電信號，再通過類比/數位轉換器轉換成數位影像信 號，然後再傳送至數位信號處理器中，經濾波處理後， 將可以辨識的信號傳送至攝影式閱讀裝置的通訊控制界 面卡（communication interface controller) ; (3)信號自攝影 式閱讀裝置的通訊控制界面卡經高速數據傳輸纜線與主 電腦裝置的通訊控制界面卡相連接，連接的通訊埠可為 串行埠、平行埠、PCMCIA(PC Memory Card International Association)或USB ; (4)接著，主電腦裝置將數位影像信 號轉換成影像陣列（digital image data matrix)文件，利用 OCR (Optical Character Recognition)技術對此影像陣列辨 識，得出文本關鍵詞（Text Headword)，輸入資料庫查詢 系統（Database Searching System)查詢得到所需之相關資 料（如解釋文字、聲音、影像等），並同時將之顯示於主電 腦裝置的顯示器（display心¥。6)上。 本發明之一優點在於利用攝影透鏡結合影像感應裝 置製成之攝影式閱讀，具有體積小、使用靈活等特點， 將其應用於印一體文字之影像輸入，便於操作。 本發明之另一個優點在於本裝置提供了印刷體文字 之線上即時辨識、即時查詢、即時顯示的功能，實現閱 本纸張尺度適用肀SS家標率（CNS > A4規格（2丨0X297公釐） I 裝 訂 咸 (請先閲讀背面之注意事項再填寫本頁)Yin Jia 4344 96, Consumer Work Cooperative, Central Bureau of Economic Affairs, Ministry of Economic Affairs V. Description of the Invention (1) --- The present invention relates to a kind of data input device, and particularly relates to a photographic smart reading device and its operation method . According to today's open society, media information is increasingly developed, and new information, especially newspapers, miscellaneous books, and books and magazines, is the most common cause of difficulties in understanding. Therefore, it is imminent to query information or interpretation of related words. According to the common methods in the past, usually a heavy dictionary is moved for manual reference to achieve the purpose of understanding. This method is time-consuming and labor-intensive. If an information material has multiple rare words, it will inevitably reduce the reading speed and distract the reader. As a result, many computer word translation software have recently appeared to replace some or all of the functions of tired word interpretation. "However, the currently developed word translation software needs to be typed through the keyboard or pen when interpreting printed words. To save the manual steps of looking up the dictionary, but added the action of typing the word vocabulary 'If you encounter a long vocabulary, the speed is not as fast as manual query, and spelling errors will inevitably occur during typing. Users often It needs to be modified, it is very frequent, it is easy to disturb the reading thinking, and affect the reading effect. In view of this, the purpose of the present invention is to overcome the contradiction between the fastness and convenience of printing of printed words, and to provide a photographic reading device by combining electronic technology and programming. This type of photographic reading device includes a two-dimensional CCD or CMOS image sensor module. Its operation steps and internal control processes are: (1) the user first passes the photographic reading to the host computer device via High-speed transmission cable connection, and then the front lens of the photographic reading device is pointed at the print to be queried. 3 This paper size is applicable to China National Standard (CNS) A4 specification (210X297). ^ 1 Order ^ (Please read the note on the back first Please fill in this page again} Printed by the Central Standards Bureau of the Ministry of Economic Affairs, Shellfish Consumer Cooperative, 4344 96 Λ7 B7____ V. Description of the invention (2) Body text or phrase, press the acquisition switch to capture the text image through the light photography lens The optical signal is simultaneously transmitted to the image sensing device (CCD or CMOS image sensing device) in the photographic reading device; (2) inside the aforementioned image sensing device, the image driver is controlled by a digital signal processor to drive The image sensor converts the optical signal into an analog electrical signal, which is then converted into a digital image signal by an analog / digital converter, and then transmitted In the digital signal processor, after filtering, the recognizable signal is transmitted to the communication interface controller of the photographic reading device; (3) The signal is transmitted from the communication control interface card of the photographic reading device through high speed. The data transmission cable is connected with the communication control interface card of the host computer device. The connected communication port can be a serial port, a parallel port, PCMCIA (PC Memory Card International Association) or USB; (4) Then, the host computer device will digitally The image signal is converted into a digital image data matrix file. The OCR (Optical Character Recognition) technology is used to identify the image array, and the text headword is obtained. Enter the database search system to query Related information (such as explanatory text, sound, video, etc.), and simultaneously display it on the display of the main computer device (display center ¥. 6). One of the advantages of the present invention lies in the use of a photographic lens combined with an image sensing device. Chengzhi's photographic reading has the characteristics of small size and flexible use. It is applied to the image input of printed text, which is easy to operate. Another advantage of the present invention is that the device provides the functions of online real-time identification, real-time query, and real-time display of printed text, so that the paper can be read. Standard rate (CNS > A4 size (2 丨 0X297mm) I Binding (Please read the precautions on the back before filling this page)
4344 -S 經濟部中央搮率局貝工消費合作社印裝 A7 五、發明説明（3 ) 讀與理解同步的過程，大幅地提升閱讀者的閱讀速度。 本發明之再一優點為利用此裝置之軟、硬體，可方 便而且即時地完成在個人電腦、手持式個人電腦 (Handheld Personal Computer)或電子字典上對印刷體文 字之自動輸入。 為了進一步說明本發明之架構、操作方式及特點， 下面就配合附圓說明本發明之較佳實施例’其中： 第1圖係繪示本發明之攝影式智慧閱讀裝置的架構 圖式； 第2圖係繪示本發明之攝影式智慧閱讀裝置的閱讀 裝置架構之方塊圖； 第3圖係繪示應用於第2圖之攝影式閱讀裝置的影 像感測器電路圖； 第4圖係繪示本發明之攝影式智慧閱讀裝置的主電 腦裝置架構之方塊圖； 第5圖係繪示本發明中攝影式閱讀裝置擷取文字影 像的流程圖； 第6圖係緣示本發明中主電腦裝置處理文字影像的 流程圖； 第7圖係緣示本發对主電腦裝置進行文字影像辨 識的流程囷； 第8圖係㈣本發财對文字影像進行框架處理的 7F意圖， 第9圖係_示在本發明的字元辨識程序中特徵像素 I ^IT.^-(請先閲讀背面之注意事項再填寫本頁)4344 -S Printed by the Shellfish Consumer Cooperative of the Central Bureau of Economic Affairs of the Ministry of Economic Affairs A7 V. Description of the invention (3) The process of synchronizing reading and understanding greatly improves the reading speed of readers. Another advantage of the present invention is that the software and hardware of the device can be used to conveniently and immediately complete the automatic input of printed characters on a personal computer, a handheld personal computer or an electronic dictionary. In order to further explain the architecture, operation mode, and characteristics of the present invention, the following describes the preferred embodiment of the present invention with the attached circles. Among them: FIG. 1 is a diagram showing the architecture of the photographic smart reading device of the present invention; FIG. Is a block diagram showing the reading device architecture of the photographic reading device of the present invention; FIG. 3 is a circuit diagram of an image sensor applied to the photographing reading device of FIG. 2; FIG. 4 is a drawing Block diagram of the main computer device structure of the invented photographic intelligent reading device; FIG. 5 is a flowchart illustrating the text image capture by the photographic reading device in the present invention; FIG. 6 is a diagram illustrating the processing of the main computer device in the present invention Flowchart of text image; Figure 7 shows the flow of text image recognition on the host computer by the camera; Figure 8 shows the 7F intention of the frame processing of the text image by the camera, and Figure 9 shows Feature pixel I ^ IT. ^-In the character recognition program of the present invention (please read the precautions on the back before filling this page)
434496 經濟部中央樣率局貝工消費合作社印装 A7 B7 立、發明説明（4) 選取方式的示意圖； 第10圖係繪示本發明的循環辨識程序中進行匹配判 別的流程圖； 第11圖係繪示對資料庫進行關鍵詞查詢動作的流程 圖。 實施例之說明 本發明利用攝影式閱讀裝置（内含一個二維CCD或 CMOS製程元件之光影像感測裝置），將書本、雜誌上之 印刷體文字影像通過攝影式閱讀裝置前端之透鏡（lens) 傳入攝影式閱讀裝置内影像感測裝置，經類比/數位（A/D) 轉換後，利用數位信號處理（DSP)技術將影像濾波，並同 時透過高速傳輸線路傳至主電腦裝置，利用〇CR技術辨 識出文本關鍵詞’輸入資料庫查詢系統查詢得到所需相 關資料（如解釋文字、聲音、影像等），並同時將之顯示於 主電腦裝置的顯示器（display device)上。 本發明之裱影式智慧閲讀裝置的硬體架構說明如 下。首先’請參閱第1圖’本發明之攝影式智慧閱讀裝 置係包括：一攝影式閱讀裝置10，作為影像攝入裝置， 用以攝入文數字、條碼、影像；一主電腦裝置12，作為 整個系統的控制核心；及一高速數據傳輸纜線14，連接 前述攝影式閱讀裝置10及主電腦裝置12，用以傳輸影 像資料。當攝影式閱讀裝置10攝入影像後，經由高速數 據傳輸纜線14將影像資料傳輸給主電腦裝置12，藉以 提供不同應用之所需。例如，當主電腦裝置12為一電子 本紙張尺度適用中國困家橾隼（CNS > A4規格（210X297公釐） (請先閲請背面之注意事項再填寫本頁) 、νβ 434496 經濟部中央標率局貞工消费合作社印製 A7 B7 五、發明説明（5) 字典，則在攝影式閱讀裝置10經高速數據傳輪纜線14 輸入字元、單字或片語的影像資料後，主電腦裝置12會 顯示出該字元、單字或片語的相關資料β 請參閱第2圖，本發明之攝影式智慧閱讀裝置的攝 影式閱讀裝置10係經由高速數據傳輸纜線14連接至通 訊控制介面30 ’再經由通訊控制介面30將資料輸入給 主電腦裝置12。前述攝影式閲讀裝置10係包括：一光 攝影透鏡16 ’作為文字影像的成像裝置；一開關μ ; 一二維影像感應器（Image sensor)20 ，例如一 CCD或 CMOS影像感測器’用以接收經前述光攝影透鏡16成像 之文字影像的光信號，並將光信號轉換成類比電信號； 一類比數位（A/D)轉換器22 ，用以將由前述影像感應器 獲得之類比信號轉換為數位信號；一數位信號處理器 (DSP)24 ’用以受前述開關18控制，可接收經前述類比 數位轉換器22輸出之數位影像訊號，對該數位影像訊號 進行遽波處理；一記憶裝置26，可為隨機存取記憶體 (RAM)及唯讀記憶體（ROM)，其中，唯讀記憶體是用以儲 存控制數位信號處理器24操作所需之程式，而隨機存取 記憶體則是用於影像感應器20及數位信號處理器24的 一般操作中。 前述攝影式閱讀裝置10可自印刷品28上擷取影 像。前述攝影式閱讀裝置10的透鏡16置於其前端，且 透鏡16的設計最好能配合使用者使用的方式，同時提供 給物端（即印刷品28)較長的景深。 本紙張尺度適用t國國家標隼（CNS) A4堤格（210X297公釐） ----------^— (請先閲讀背面之注意事項#填寫本頁) 訂 4 A7 B7 6 五、發明説明（ 前述高速數據傳輸缆線14可為一平行埠傳輸獲線， 用以連接至主電腦裝置的平行皡。或可為一串行蜂傳輸 纜線，用以連接至主電腦裝置的串行埠。或者，亦可為 符合通用串列匯流排（universal serial bus，簡稱USB)協 定’或 PCMCIA(PC Memory Card International Association) 或PC卡標準的.通訊方式之镜線。當然，傳輪的方式若是 為了要避免使用纜線’也可以採用無線或是紅外線的方 式β 前述開關18可為一即時開關，也就是當使用者按著 開關18時’攝影式閱讀裝置10便持續地讀入影像，而 在使用者放開開關18時’攝取一靜態影像。或者，開關 18亦可為一般的按鍵開關’當開關μ開啟時，便攝取影 像’當開關18關閉時，便停止攝取影像9 前述影像感應器20係受一影像感測驅動裝置控制， 以將擷取到的影像轉換成類比影像訊號。而在本實施例 中’前述影像感測驅動裴置係儲存於記憶裝置26中的程 式指令，此程式指令同時亦控制前述數位信號處理器24 的運作。 前述數位仏號處理器24為整個攝影式閱讀裝置的控 制核心，在本實施例中可為微處理器、微電腦、微控制 器或RISC處理器等’其主要動作如下：（1)受開關丨^控 制接收影像訊號；（2)對類比數位轉換器22輸出之數位影 像信號進行濾波處理，即去除信號之失真部份，同時進 行影像信號之邊緣定位，結構加強及平滑處理，藉以得 木紙張尺度適用中國國家橾準（CNS } A4規格（2丨0X297公着 ---------^------ir------^ (請先聞讀背面之注意事項再填寫本頁) 經濟部中央標準局—工消費合作社印裝 434496 經濟部中央標準局貝工消费合作社印笨 A7 B7 五、發明説明（7) 到較為理想的數位影像信號；（3)將數位影像信號轉換為 主電腦裝置可識別之圖像數位信號，傳至通信控制介面 30 - 請參閱第3圖’前述影像感應器（image sensor)係包 括.一影像感測電路34，及一移位暫存器（CCD analog shift register)44 ,經由一移位閘 42(shift gate)與前述影像 感測電路34連接。前述影像感測電路34係包括排列成 一維矩陣型式的複數個光感測元件，且每一個光感測元 件係包括：並聯的一光電二極管（ph〇t〇doide)36及一電容 器38，且其一端接地40。其具體運作方式為：光影像 信號傳入光電二極管36後，經其光電感應過程轉變為一 信號電荷（signal charge)，並將之存於電容器38中，再經 移位閘42傳至移位暫存器44中，轉換為類比影像信號 (analog image signal)。前述光電二極管36可用一光阻器 (photo resistor)取代，而移位暫存器44亦可採用CM〇s 製程之類比移位暫存器。 請參閱第4圖，本發明之攝影式智慧閱讀裝置的主 電腦裝置12係包括：一處理單元46，為整個系統的控 制核心，可為CISC或RISC處理器，用以執行〇CR文字 辨識命令及字彙查詢命令；一顯示裝置48，受前述處理 單元46控制，藉以顯示自該處理單元46輸出的資訊。 前述處理單元46經由匯流排50連接至至少一記憶裝置 52，至少一儲存裝置54，及通訊控制介面3〇。 前述記憶裝置52可包括隨機存取記憶體及唯讀記憶 ----------¢------1T------^ (請先閱讀背面之注意事項再填寫本頁) 434496 A 7 B7 經濟部中央輮準局貝工消费合作社印«. 五、發明説明（8) 體。而前述儲存裝置54則依據主電腦裝置的不同型式， 而可為軟碟 '硬碟、光碟、記憶體或PCMCIA記憶卡等》 例如，在一電子字典中，單字及片語的定義及相關資訊 便可儲存於儲存裝置54中。在本發明中，影像辨識及資 料庫查詢的應用程式也可儲存在儲存裝置54中。前述通 訊控制界面30，可經由平行埠（parallel port)、串行埠 (serial port)或通用串列匯流排（universal serialbus ’ 簡稱 USB)連接於前述高速數據傳輸纜線14。 請參閱第5圖，其係繪示攝影式閱讀裝置擷取影像 的流程，包括下列步驟：（1)步驟60 ’決定使用者是否將 攝影式閱讀裝置對準待辨識之印刷髅文字或片語’按動 採集開關，若無則回到啟始，否則進行下一步驟；（2)步 驟62，經由光攝影透鏡16擷取影像的光信號，同時經 由影像感應器20產生矩陣格式的類比影像信號；（3)步驟 64，前述類比影像信號通過類比/數位轉換器22轉換成 數位影像信號，然後再傳送至數位信號處理器24中；（4) 步驟66，經數位信號處理器對數位影像信號進行濾波處 理及對應的信號轉換，並將處理後之影像資料陣列（image data matrix)傳送至攝影式閱讀裝置的通訊控制界面 (communication interface controller) ; (5)步驟 68，攝影 式閱讀裝置的通訊控制界面將影像資料陣列經高速數據 傳輸纜線傳輸至主電腦裝置的通訊控制界面，連接通訊 控制界面的通訊埠可為串行埠、平行埠、PCMCIA(PC Memory Card International Association)或 USB 等，然後 ίο (請先閲讀背面之注意事項再填寫本頁) .裝- 本紙張尺度適用中國國家標準（CNS ) Α4規格（210Χ297公釐） 經濟部中央橾率爲貝工消费合作社印装 434496 A7 B7 ______ 五、發明説明（9 ) 再回到啟始，決定是否再擷取影像。 請參閱第6圖，其繪示主電腦裝置的操作流程’包 括下列步驟：（1)步驟70，決定主電腦裝置是否自攝影式 閱讀裝置讀取影像資料陣列，若無則回到啟始’否則進 行下一步驟；（2)步驟72，處理單元開始進行字元辨識程 序；（3)步驟74，進行資料庫查詢程序；（4)步驟76，依 據查詢結果，自資料庫中擷取相關資訊；（5)步驟78 ’將 查詢得到所需之相關資訊（如解釋文字、聲音、影像等） 顯示於主電腦孝置的顯示器（display device)上。 請參閱第7圖，前述字元辨識程序的流程主要是包 括下列步驟··（1)步驟80，經由攝影式閱讀裝置讀取影像 資料，並將其以陣列的形式儲存於主電腦裝置的RAM 申；（2)步驟82 ，將影像資料陣列之多級灰度格式 (multi-gray format)或彩色格式（color image format)轉化 成二級灰度格式（binary-gray format) ’即以0與1表示像 素之有無，藉以區分文字與背景影像，以便於影像辨識； (3)步驟84，根據各像素之稀疏綢密特徵，進行二值化的 影像陣列所具有的雜訊過濾，去除一些特別稀疏的雜 點，使待辨識之影像點陣整齊、清晰；（4)步驟86，將點 陣影像依據其點陣疏密（point density)，將其區分為多塊 片語影像（word image blocks)和多塊字元影像（character image blocks) ’以便於單個字元的辨識；（5)步称88，進 行字元或片語影_像循環辨識程序；（6)步驟，決定是否 辨識下一個字元或片語，若是則回到啟始，否則辨 11 本紙目轉奉（CNS) ( 210x297公着)" ----- n n - n I ! I ^ 旅 (請先聞讀背面之注意事項再填寫本頁) 經濟部中央梯準局貝工消費合作社印^ 4 3^4 96 A7 B7 Γ — - ...... " ------- 五、發明説明（1Q ) 識。 在前述步驟80中’由攝影式閱讀裝置輪入主電腦裝 置的影像資料陣列，格式函數為Pmn(Xm Yn)，其中，（Xm Yn) 為像素的座標位置，Pmn(Xm，Yn)為該像素的影像多階灰 度值。434496 Printed A7 B7 by the Central Sample Rate Bureau of the Ministry of Economic Affairs, Shellfish Consumer Cooperative Co., Ltd. (4) Schematic diagram of the selection method; Figure 10 is a flowchart showing the matching discrimination in the cycle identification procedure of the present invention; Figure 11 It is a flow chart showing the keyword query action on the database. Description of the embodiment The present invention uses a photographic reading device (a light image sensing device containing a two-dimensional CCD or CMOS process element) to pass printed text images on books and magazines through a lens at the front end of the photographic reading device ( lens) is introduced into the image sensing device in the photographic reading device, and after analog / digital (A / D) conversion, the digital image processing (DSP) technology is used to filter the image, and at the same time, it is transmitted to the host computer device through the high-speed transmission line. Use 〇CR technology to identify the text keywords' input database query system to get the required relevant information (such as explanatory text, sound, video, etc.), and display it on the display device of the main computer device at the same time. The hardware architecture of the shadow-mounted smart reading device of the present invention is described below. First, please refer to FIG. 1. The photographic smart reading device of the present invention includes: a photographic reading device 10 as an image ingesting device for ingesting characters, barcodes, and images; and a host computer device 12 as The control core of the entire system; and a high-speed data transmission cable 14 connected to the aforementioned photographic reading device 10 and the host computer device 12 for transmitting image data. After the photographic reading device 10 takes in the images, the image data is transmitted to the host computer device 12 via the high-speed data transmission cable 14 so as to provide the requirements for different applications. For example, when the main computer device 12 is an electronic paper, it is suitable for Chinese families (CNS > A4 size (210X297 mm) (please read the precautions on the back before filling this page)), νβ 434496 Central A7 B7 is printed by the Standards Bureau Zhengong Consumer Cooperative. 5. Description of the invention (5) Dictionary. After the photographic reading device 10 enters the image data of characters, words or phrases through the high-speed data transmission cable 14, the host computer The device 12 will display the relevant information of the character, word or phrase β Please refer to FIG. 2. The photographic reading device 10 of the photographic smart reading device of the present invention is connected to the communication control interface via a high-speed data transmission cable 14. 30 ', and then input data to the host computer device 12 through the communication control interface 30. The aforementioned photographic reading device 10 includes: a light photographing lens 16' as an imaging device for text images; a switch μ; a two-dimensional image sensor ( Image sensor) 20, for example, a CCD or CMOS image sensor is used to receive the optical signal of the text image imaged by the aforementioned optical photographic lens 16 and convert the optical signal into an analog electrical signal An analog-to-digital (A / D) converter 22 for converting analog signals obtained by the aforementioned image sensor into digital signals; a digital signal processor (DSP) 24 'for controlling by the aforementioned switch 18 and receiving The digital image signal output by the aforementioned analog digital converter 22 performs wave processing on the digital image signal; a memory device 26 may be a random access memory (RAM) and a read-only memory (ROM), of which, the read-only The memory is used to store programs needed to control the operation of the digital signal processor 24, and the random access memory is used for the general operation of the image sensor 20 and the digital signal processor 24. The aforementioned photographic reading device 10 may Capture an image from the printed matter 28. The lens 16 of the aforementioned photographic reading device 10 is placed at the front end, and the design of the lens 16 should preferably match the way the user uses it, and at the same time provide a longer object (ie, printed matter 28). Depth of field. This paper size is applicable to national standard (CNS) A4 (210X297 mm) ---------- ^-(Please read the notes on the back first # Fill this page) Order 4 A7 B7 6 V. Description of the invention ( The aforementioned high-speed data transmission cable 14 may be a parallel port transmission and acquisition line for connecting to a parallel port of the host computer device, or may be a serial bee transmission cable for connection to a serial port of the host computer device. Alternatively, it can also be a mirror line of the communication method that conforms to the universal serial bus (USB) protocol 'or the PCMCIA (PC Memory Card International Association) or PC Card standard. Of course, if the transmission method is intended for To avoid using cables, you can also use wireless or infrared. The aforementioned switch 18 can be an instant switch, that is, when the user presses the switch 18, the 'photographic reading device 10 continuously reads in images, and When the user releases the switch 18, 'take a still image. Alternatively, the switch 18 can also be a general key switch 'When the switch μ is turned on, the image is captured' When the switch 18 is turned off, the image capture is stopped 9 The aforementioned image sensor 20 is controlled by an image sensing driving device to The captured image is converted into an analog image signal. In this embodiment, the aforementioned image sensing driving program is a program instruction stored in the memory device 26, and this program instruction also controls the operation of the aforementioned digital signal processor 24 at the same time. The aforementioned digital tweeter processor 24 is the control core of the entire photographic reading device. In this embodiment, it can be a microprocessor, a microcomputer, a microcontroller, or a RISC processor. Its main actions are as follows: (1) Received by the switch 丨^ Control the received image signal; (2) Filter the digital image signal output by the analog-to-digital converter 22, that is, remove the distortion part of the signal, and at the same time perform edge positioning, structure strengthening and smooth processing of the image signal to obtain wood paper. Standards are applicable to China National Standards (CNS) A4 specifications (2 丨 0X297) --------- ^ ------ ir ------ ^ (Please read the notes on the back first Refill this page) Central Standards Bureau of the Ministry of Economic Affairs—Industrial and Consumer Cooperative Cooperative Printing 434496 Central Standards Bureau of the Ministry of Economic Affairs and Industrial Cooperative Cooperative Printing Ben A7 B7 V. Description of the invention (7) To the ideal digital image signal; (3) Digital The image signal is converted into a digital image signal that can be recognized by the main computer device and transmitted to the communication control interface 30-please refer to FIG. 3 'the aforementioned image sensor includes: an image sensing circuit 34, and a shift Register (CCD ana log shift register 44, which is connected to the image sensing circuit 34 via a shift gate 42. The image sensing circuit 34 includes a plurality of light sensing elements arranged in a one-dimensional matrix, and each light The sensing element includes: a photodiode 36 and a capacitor 38 connected in parallel, and one end of the capacitor is grounded to 40. The specific operation mode is as follows: after the light image signal passes through the photodiode 36, it is photoelectrically sensed by the photodiode 36 The process is transformed into a signal charge, which is stored in the capacitor 38, and then transferred to the shift register 44 via the shift gate 42 to be converted into an analog image signal. The aforementioned photodiode 36 may be replaced by a photo resistor, and the shift register 44 may also adopt an analog shift register of the CMOS process. Please refer to FIG. 4, the master of the photographic smart reading device of the present invention The computer device 12 includes: a processing unit 46, which is the control core of the entire system, and can be a CISC or RISC processor for executing the OCR text recognition command and the vocabulary query command; a display device 48, subject to the foregoing processing The unit 46 controls to display the information output from the processing unit 46. The aforementioned processing unit 46 is connected to at least one memory device 52, at least one storage device 54, and a communication control interface 30 via a bus 50. The aforementioned memory device 52 may include Random Access Memory and Read Only Memory ---------- ¢ ------ 1T ------ ^ (Please read the precautions on the back before filling this page) 434496 A 7 B7 Printed by the Shellfish Consumer Cooperative of the Central Bureau of Standards and Quarantine of the Ministry of Economic Affairs. V. Invention Description (8). The aforementioned storage device 54 may be a floppy disk, a hard disk, a CD-ROM, a memory, or a PCMCIA memory card according to different types of the host computer device. For example, in an electronic dictionary, the definition of words and phrases and related information It can be stored in the storage device 54. In the present invention, the application programs of image recognition and database query can also be stored in the storage device 54. The communication control interface 30 can be connected to the high-speed data transmission cable 14 through a parallel port, a serial port, or a universal serial bus (USB for short). Please refer to FIG. 5, which illustrates the process of capturing images by the photographic reading device, including the following steps: (1) Step 60 'Decide whether the user will point the photographic reading device at the printed cross-word or phrase to be recognized. 'Press the acquisition switch, if not, go back to the beginning, otherwise proceed to the next step; (2) Step 62, capture the optical signal of the image through the optical photography lens 16 and generate an analog image in matrix format through the image sensor 20 Signal; (3) step 64, the aforementioned analog image signal is converted into a digital image signal by the analog / digital converter 22, and then transmitted to the digital signal processor 24; (4) step 66, the digital image is processed by the digital signal processor The signal is subjected to filtering processing and corresponding signal conversion, and the processed image data matrix is transmitted to the communication interface controller of the photographic reading device; (5) Step 68, The communication control interface transmits the image data array to the communication control interface of the host computer device through a high-speed data transmission cable, and connects to the communication control industry. The communication port on the front can be a serial port, a parallel port, PCMCIA (PC Memory Card International Association) or USB, etc., and then (Please read the precautions on the back before filling out this page). Installation-This paper size applies to Chinese national standards (CNS) Α4 size (210 × 297 mm) The central government department of the Ministry of Economic Affairs printed 434496 A7 B7 for Shellfish Consumer Cooperatives. 5. Description of the invention (9) Return to the beginning and decide whether to capture images again. Please refer to FIG. 6, which shows the operation flow of the host computer device 'including the following steps: (1) Step 70, decide whether the host computer device reads the image data array from the photographic reading device, and if not, return to the beginning' Otherwise, proceed to the next step; (2) step 72, the processing unit starts the character recognition process; (3) step 74, the database query process; (4) step 76, according to the query result, retrieve the correlation from the database Information; (5) Step 78 'Display the relevant information (such as explanatory text, sound, image, etc.) required by the query on the display device of the host computer. Please refer to FIG. 7. The flow of the character recognition program mainly includes the following steps: (1) Step 80: Read image data through a photographic reading device and store it in the form of an array in the RAM of the host computer device. (2) Step 82: Convert the multi-gray format or color image format of the image data array into a binary-gray format. 1 indicates the presence or absence of pixels, so as to distinguish between text and background images to facilitate image recognition; (3) step 84, according to the sparse and dense characteristics of each pixel, perform a noise filtering of the binary image array to remove some special The sparse noise makes the dot matrix of the image to be identified neat and clear; (4) Step 86, the dot matrix image is divided into word image blocks according to its point density. ) And multiple character image blocks ('character image blocks') to facilitate the identification of a single character; (5) Step 88, which performs a character or phrase image_image loop recognition process; (6) Step, decide whether to recognize One character Phrase, if it is, go back to the beginning, otherwise discern 11 papers (CNS) (210x297) " ----- nn-n I! I ^ Brigade (Please read the notes on the back before reading (Fill in this page) Printed by the Central Laboratories of the Ministry of Economic Affairs, Shellfish Consumer Cooperatives ^ 4 3 ^ 4 96 A7 B7 Γ —-...... " ------- 5. Description of Invention (1Q). In the foregoing step 80, the image data array of the host computer device is rotated by the photographic reading device, and the format function is Pmn (Xm Yn), where (Xm Yn) is the coordinate position of the pixel, and Pmn (Xm, Yn) is the The image's multi-level grayscale value.
PiiOi，yi) P2iCx2,yi) P3l(X3,yi) Pftn- Ϊ、1 (Xfn-1，Y i) P_ ιίχ Vi、 Pl2(Xl,Y2) P22(X2，y2) P32(X3，ya p v-i^ * · · · » Pln(Xl,yn) P2n(X2,yn) P3n(X3,yn) P(m-1 }(π· 1 )(Xm，1，yn-l) Pmn(xm，yn) 在前述步驟86中，經過濾波處理過的影像資料陣 列，利用影像分離與連續之特徵，切割成各個獨立字元 的影像’此即為“框架處理，，(framingpr〇ces〇。 各個獨立之字元影像陣列分別以gi，以，^ ^ 表示，請參閱第8圖’以“BIZ”為例，其經切分後，“B” 的影像區塊以g,表示，“I”的影像區塊以g2表示，而“z„ 的影像區塊以g3表示。函數R(G)為影像循環辨識函數， 將m個獨立影像塊gl、g2、g3 ‘ ..、心代入R(G)得 R(gi、g2、g3 . . .、gt)，可辨識得相對應的字元ei、 e2、以.· ·、et ’循序儲存於數組E中，數組E即為 辨識所得之關鍵詞。 前述影像循環辨識函數R(G)所採用的字元辨識方法 基本上為匹配判別。請參閱第9圖，以印刷體的“Z”為例， 其具有的筆段（stroke)為S-1 92、S-2 94及S-3 96 ,首先 12 ----------装------iT------Λ (請先閲讀背面之注意事項再填寫本頁) 本紙張尺度適用十國國家榡準{〇^/114说格（2〗〇\297公釐） 經濟部中央標率局貝工消費合作社印製 434496 A7 ____B7 五、發明説明（u ) 決定每-筆段的相對長度，在本财，S 2 94較長，而 S-192及S-396則等長;接著決定字元影像的筆段數目， 以本例而言，共有三個筆段；決定每—筆段的方向在 本例中，S-192及S-3 96均為水平方向，s_2 94為右上 至左下的方向；最後再決定字元影像的連接點數目，在 本例中為一。然後，根據上述步驟所得到的特徵資訊， 於資料庫查詢的程序中，再與字元特徵庫中的每一個字 元的特徵減較。前述字元賴料預先时於資料庫 中，所以本發明之字元辨識，依據内建範本不同，除了 可辨識各種不同的文字、數字外，尚可應用於辨識影像。 為了更進一步說明影像循環辨識，請參閱第1〇園， 其缘示在步驟88中时元或片語影像循環辨識程序的流 程’包括下列步驟⑴步驟⑽’識別字元影像的筆段； (2)步驟102，決定字元影像的筆段數目；⑺步驟1〇3， 決定每-筆段_對長度；⑷㈣⑽，決定相連接筆 段間的角度U5)㈣105，決定每一筆段的方向；（6)步 驟106 ’根據上述步驟取得字元的特徵資訊；⑺步驟 108,將每-個字元像素的特徵與字元特徵庫中的每一個 字元影像的特徵相比：(8)步称110，是否找到特徵相匹 配的字元，若是則進行至步称112，否則進行至步称114; ⑼步驟112，使用在子元特徵庫中找到的字元做為參考 字70碼；（10)步驟114，使用在字元特徵庫中找到特徵最 接近的字元做為參考字元碼。 請參閱第11圖，其繪示對關鍵詞進行查詢動作的流 13 本紙張从逍用帽@家揉準（CNS)舢祕（21gx297公廣） ----------^------ΐτ------4. (諳先閲讀背面之注意事項再填寫本頁) 五、發明説明（12 程’包括下列步驟：（I)步騍 識轉換後的文本關鍵詞;(2)步驟；i8 經辨 文本關：詞與資料庫中既有的資料進行—比較資= 120，蚊是否找到相符的資料，若是，則進行至步: 124，否則進行至步驟126 ;⑷步驟12 ::步驟 取此關鍵詞的相關資訊；（5)步 料庫中獲 畏接折的相關咨π . 止 獲取與此關鍵詞 ! 50’將獲取的相關資訊顯 不於液晶顯示器上。 只叫湖 前述步驟250中，在杳詢眘姐虑& 資降了宝分資料庫後’所獲取的相關 資訊除了予凡、早子或片語的定義與說明外， 侷限於文字的資料，而是可包括 1 媒體資料。 子tw象、聲音等多 在上述的實施例中，主電腦裝置可為任何包含有微 處理器的裝置’如個人電蹈、手持式個人電腦或電子 字典，例如英業達股份有限公司生產的無敵CD 腦閱讀辭典等。 I 裝------Λ1Τ------^ (請先閲讀背面之注項再填寫本頁) 經濟部中央標準局貝工消費合作社印装 本紙張尺度適用中國固家標率（CNS )八4規格（210X297公釐）PiiOi, yi) P2iCx2, yi) P3l (X3, yi) Pftn- Ϊ, 1 (Xfn-1, Y i) P_ ι Vi, Pl2 (Xl, Y2) P22 (X2, y2) P32 (X3, ya p vi ^ * · · »Pln (Xl, yn) P2n (X2, yn) P3n (X3, yn) P (m-1) (π · 1) (Xm, 1, yn-l) Pmn (xm, yn) In the aforementioned step 86, the filtered image data array uses the features of image separation and continuity to cut the image into individual characters. This is "framing processing," (framingpr〇ces〇. Each independent character The meta image array is represented by gi, and ^ ^, please refer to Figure 8 'Take "BIZ" as an example. After segmentation, the image block of "B" is represented by g, and the image area of "I" The block is represented by g2, and the image block of "z" is represented by g3. The function R (G) is the image loop identification function. Substituting m independent image blocks gl, g2, g3 '.., into R (G), we get R (gi, g2, g3..., Gt) can be identified as the corresponding characters ei, e2, and stored in the array E in the order of .., et '. The array E is the identified keywords. The character recognition method used in the aforementioned image loop recognition function R (G) is basically Matching discrimination. Please refer to Fig. 9. Taking the "Z" of the printed body as an example, it has strokes of S-1 92, S-2 94, and S-3 96. First, 12 ----- ----- 装 ------ iT ------ Λ (Please read the notes on the back before filling out this page) This paper size is applicable to the ten countries' standards {〇 ^ / 114 说 格 ( 2 〖〇 \ 297mm） Printed by the Central Standards Bureau of the Ministry of Economic Affairs, Shellfish Consumer Cooperative, 434496 A7 ____B7 V. Invention Description (u) Determines the relative length of each segment. In this case, S 2 94 is longer, and S-192 and S-396 are equal in length; then determine the number of strokes of the character image. For this example, there are three strokes; determine the direction of each stroke. In this example, S-192 and S -3 96 are horizontal, and s_2 94 is from top right to bottom left. Finally, the number of character image connection points is determined, which is one in this example. Then, according to the feature information obtained in the above steps, query in the database In the program, the features of each character in the character feature database are subtracted. The foregoing character relies on the database in advance, so the character recognition of the present invention is based on different built-in templates, except that In addition to recognizing various characters and numbers, it can also be used for recognizing images. For further explanation of image cycle recognition, please refer to Garden 10, whose reason is shown in step 88. The process of the cycle or phrase image cycle recognition program 'Include the following steps⑴Step⑽' to identify the strokes of the character image; (2) Step 102, determine the number of strokes of the character image; ⑺Step 103, determine the length of each-stroke_pair; ⑷㈣⑽, determine the phase Connect the angles between the pen segments U5) ㈣105 to determine the direction of each pen segment; (6) Step 106 'Get the character information of the character according to the above steps; ⑺Step 108, the feature of each-character pixel and the character The features of each character image in the database are compared: (8) step 110, whether characters with matching features are found, if yes, proceed to step 112, otherwise proceed to step 114; ⑼step 112, used in The character found in the sub-feature library is used as the reference character 70 code; (10) Step 114: Use the character with the closest feature found in the character feature database as the reference character code. Please refer to Figure 11, which shows the flow of query action on keywords. 13 papers from Xiaoyong Cap @ 家 瘦 准 （CNS） 舢 秘 (21gx297 公 广) ---------- ^- ----- ΐτ ------ 4. (谙 Please read the notes on the back before filling this page) 5. Description of the invention (12 procedures' includes the following steps: (I) Steps to identify the key of the converted text Words; (2) step; i8 pass through the discerned text: the words are compared with the existing data in the database—comparison information = 120, whether the mosquito finds matching data, if yes, go to step: 124; otherwise, go to step 126 ; ⑷Step 12 :: Steps to get the relevant information of this keyword; (5) Relevant information from the step library. Stop getting this keyword! 50 'will not show the relevant information on the LCD In the above-mentioned step 250 of the lake only, after consulting Sister Shen's consideration & downgrading the Baofen database, the relevant information obtained was limited to the text except for the definition and description of Yufan, Zaozi or the phrase. Data, but can include 1 media data. In the above embodiment, the main computer device can be any A processor device such as a personal computer, a handheld personal computer, or an electronic dictionary, such as the invincible CD brain reading dictionary produced by Inventec, Inc. I installed ------ Λ1Τ ------ ^ (Please read the note on the back before filling out this page) The paper size of the printed paper of the Central Standards Bureau of the Ministry of Economic Affairs, Shellfish Consumer Cooperative, is applicable to the China National Standard (CNS) Standard 8 (210X297 mm)