Research And Implementation Of Searching Test System Based On Image Recognition

With the continuous development of the Internet, not only in the life, work, etc., significant changes also have taken place in education field. Online education is the integration of education and the Internet, and it is a new mode of education. Online question bank is an important part of online education. We can not only search for a variety of test questions and their answers and analysis, but form a test paper through the online question bank. This has brought a great help to the teaching and learning.However,C urrently online question bank only supports the manual input,what reduces its search efficiency. For example, the title of question we need to search is long and it cannot be copied and pasted, at this time we must manually input it, although we can enter a part of it to search, but the search accuracy is low. So in this case, the search efficiency is lower.In view of the problem of online question bank, this paper presents a question search system based on image recognition technology. This system can use a picture to search a question, while the question is in this picture. In other words, we can put a question to a picture by taking a picture or a screenshot, when the title of question is long and it cannot be copied and pasted, and then we use this picture to search the question.it is very convenient for mobile phone users if the system has this function. Because they only need to use the mobile phone to take pictures to get the answers and analysis of questions.The purpose of this paper is to implement a system that can use a picture to search the question, and it is an open source and stable API system. It can provide a basic function for online question bank or other systems. This system can overcome the shortcomings of the low efficiency of the online question bank. It complies with the requirements of lifelong education and continuous integration of the Internet and education. And it provides users with a convenient and quick way to learn. The implementation of this system mainly involves the knowledge of image processing, image recognition, machine learning and so on.In this paper, the design and implementation of a search system based on image recognition is described in detail, and introduces its implementation method. The functional principle of the system is that, after the server receives the Web request, the first business is image processing, second is the image segmentation and character recognition, we can get a text information by character recognition, Finally, the text information is composed of keywords to search, and then we get the answer and analysis of the question.This paper uses django as Web framework, image processing and image segmentation are realized with Opencv library and Python's Image library, and it uses open source Tesseract- OCR engine for character recognition. For improving the correct rate of character recognition, this paper uses a method related character library training, finally it gets a good effect. And this paper chooses the SQLite as the database.Finally, the paper uses Nginx as Web server testing this system, test results show that the system achieves the desired effect.