Ocr technology can't recognize bank draft.

Brief introduction of 1.OCR bank bill identification products

OCR Bank Bill Identification Development Kit (SDK) is a general data batch collection component product, which is suitable for forms, questionnaires, documents and invoices with tabular characteristics. Through scanning, image processing, automatic classification and OCR (optical character recognition) technology, the data information in table images is accurately, quickly and truly extracted and saved, and the data results are converted into standard data formats, such as XML and Excel, and compared with the data of enterprises and institutions. This product supports handwriting and printing to recognize Chinese and English characters, numbers and symbols.

Second, the application background of OCR bank bill recognition

The traditional manual data entry method, which is also the most widely used method at present, is that the data is entered into the computer one by one by a special data entry clerk, and the data flows to the enterprise ERP system and Excel record form. This method consumes a lot of manpower and time, and the computer is only a data carrier in this scheme, but it has not been effectively applied. The disadvantages are obvious:

1. The recording task is heavy and tense, and the typist is prone to fatigue and mistakes in the state of high-intensity repetitive work;

2. Increasing the number of video recorders and computer equipment will increase the office cost;

3. The inefficient collection of basic data leads to the lag of comprehensive statistical data, which leads to the performance of other business information management systems of the company (such as ERP and CRM) being greatly reduced and affects the correct decision-making of enterprises.

The OCR technology is used to realize the automatic entry of bills, and the data of forms and bills are collected by batch scanning and identification, which greatly improves the efficiency of data collection, hands over heavy and repetitive work to the computer for processing, and gives full play to the advantages of computer processing technology. The change of bill recording subject has brought about the essential improvement of bill recording efficiency. For example, an A4 order has 15 pieces of data, and skilled operators need 120 seconds to record the order, while OCR technology only needs10.5 seconds, and the speed increase is 120S/ 1.5S = 79 times, which has obvious advantages.

Operating environment: mobile phone nave4e7.8.4 Baidu 12.25.0.

Recognition performance parameters of OCR bank bill recognition system;

1, identifying the character type.

_2; Print: Chinese (simplified and traditional), English, numbers.

_2; Handwriting: Chinese (simplified and traditional), English and numbers.

_2; Symbol category: identification code, barcode, check box, _, _, ○, ● etc. Symbols can be customized.

_2; 8 customizable characters: Japanese, Korean, Mongolian, Tibetan, Uygur, Kazakh, Kirgiz and Arabic.

2. Identify speed

_2; The recognition speed of a single table is about 1 s;

3. Recognition rate

_2; For the recognition of printed characters with neat layout, the correct rate is 99.96%.

_2; For the recognition of standard handwritten numerals and English, the correct rate is 98.5%.

_2; For standardized handwritten Chinese character recognition, the correct rate is about 85%;