Key Techniques Of Electronic Whiteboard System Based On Machine Vision

With the development of technology, more and more digital tools have been applied in education and business fields. Instead of the ordinary board, electronic white board is a new application which can store the content on the board as well as processing real-time communication. Among many kinds of technology of orientation in electronic white board, the machine vision technology is most popular, as it is precise in orientation, well-adapted, fast in recognition and low cost. However, with the surface augmentation, it is a challenge to reach high precision and fast recognition. Therefore, the target of the thesis is to study the critical technology of the electronic white board system based on machine vision.The thesis mainly studies the key technologies in the image acquisition, the image processing and the optimization of the real-time. The major work of the thesis is as follows:(1)The analysis of configuration of the CMOS imaging sensor in the infrared camera system:Through analyzing and comparing the whole frame output with the effective image output, according to the character of the image, the configuration of effective image output is confirmed, which can improve the real-time performance of the system.(2)The study of inner structure of CPLD:Through analyzing, the inner structure of CPLD is designed in modules. It mainly includes the high-effective image acquisition part to realize the CMOS imaging sensor’s effective image output configuration and the switch part to change the connection of memory for better real-time performance.(3)The exploitation of MCU’s software:There are three aspects, which are calculation of orientation in the image of single camera, the method of united computing and a new algorithm of coordinate mapping.(4) The optimization of real-time performance:In order to improve the system’s real-time performance, the MCU’s code is optimized, and a method of coordinate interpolation is proposed. (5)The system is transplanted to the SoC platform. The system based on MCU is transplanted to the prototype SoC focused on OR1200. The work mainly includes the modification of the hardware and processing of some critical problems in the software.The high-effective image acquisition, a new algorithm of coordinate mapping and the optimization of the system are the key points in this thesis, because they have close relationship with the precision and the real-time performance, which are the two key targets of the system. The experimental results on the prototype indicate:the recognition error can be limited in4mm, and the system can send90coordinates per second. This work fully meets all the design targets.