摘要:
本文提出了一種復雜表格識別和處理方法。該方法首先輸入一張未填有用信息的空表格作背景信息表格,然后輸入已填有用信息的同類表格作有用信息表格。對有用信息表格圖象采用幾何位置配準使兩幅表格的背景信息重合,然后利用數(shù)字邏輯運算和智能相關處理技術以達到對殘留背景信息的徹底清除,保留后填入的有用信息供識別處理。此方法對表格形式不限,適用于背景信息和有用信息采用同色或異色筆填寫。方法僅需第一次輸入一張空表格后,就可連續(xù)處理同類型表格。文中給出了表格處理結果。
Abstract:
A recognition and processing method of document form background information is presented. First, an original form, called Background Information Form Image (BIF1), is scanned. The same kind of forms, which is filled by various characters that one wants to recognize and process, called Useful Information Form Image (UIF1), is stored. Then, a geometric remedy of UIFI is used for the best matching with the BIFI. A digital logical operation is carried out in order to erase the most of background information and an intelligent correlation technique is available to completely delete the remained background information. This method is suitable for processing various forms, including background and useful information printed or filled with same colour. It can continuously process same kind of UIFIs with the input of BIFI only once. The document forms are shown and satisfactory results are obtained.
下載:
