Application Of Improved Hough Transform In Certificate Information Collecting System

With the acceleration of urbanization process in China, population migration and mobility has become a fairly common phenomenon. It’s particularly important for personal identity and information of all kinds of inquiries and identification. Documents of various kinds as a symbol of people’s identity information, they are unified security management has become an important issue of social security.As present, with the rapid development of automated identification technology and digital image processing technology, documents of various kinds of information collected to identify the product come into being. These products have entered the era of integration, networking and information technology. Especially in the aviation, banking, public security, restaurants, schools, customs, hospitals, Internet cafes and public institutions such industries, they need various types of documents and notes fast extraction and accurate identification, which is also based on information technology and security management. So, how to make key information on various types of documents fast and accurate extraction and unification of management has become a top priority at this stage. This stage documents information collection system used in a wide range of technology includes optical character recognition (OCR) and radio frequency tag identification (RFID). Only the second generation ID card and a small number of other card use smart card technology. The interior including a save individual images and information on RFID chips can be read using RFID technology. Other documents such as passports, driver’s license, passes, tickets, receipts, student card and did not use the RFID technology. Documents’optical character recognition is automatic identification technology indispensable.This paper mainly presents the following completed work:1. Describes documents information acquisition system development, application status and structural composition in detail. For the characteristics of the system, the paper propose document image distortion correction, brightness compensation, image smoothing and maximum between-class variance method series image pre-processing algorithm processes,2. Analyzes the limitations of the basic principles of the standard Hough transform and its applications. Design an improved Hough algorithm based on the OpenCV contours scanner. The algorithm can quickly and accurately navigate to the document image region.3. VC++6.0compiler platform for software, combine Intel open source computer vision library OpenCV for program development.Practice has proved that the algorithm is able to overcome the influence of the background and noise, and accurate and fast extraction to document image area in real time. It has laid a good foundation for subsequent extraction and recognition.