Research Of Educational Image Retrieval Based On Hadoop

With the rapid development of Internet and multimedia technologies,the number of images has been rapidly increasing and has been widely used in various industries.In the field of education,the image has become an indispensable material in the teaching process because of its vivid and easy-to-understand characteristics.With the development of Educational Informatization,the application of cloud computing technology in education has gradually deepened,the storage of mass resources has been solved effectively,the co-construction and sharing of teaching resources have been realized,and the speed of resource growth has been further accelerated.At the same time,the rapid expansion of the image resources has also brought difficulties to the image retrieval.In this context,this article combined with content-based image retrieval technology and Hadoop platform,proposes an image retrieval strategy that is more suitable for teaching resource platform.The strategy can improve the retrieval performance,improve the functionality of the resource platform,increase the utilization of the image resources.It plays an important role in the construction of teaching resources and the advancement of Educational Informationization.This article first summarizes the development of image retrieval technology and cloud computing technology,and analyzes the current status of educational image resource retrieval,and expounds the significance and feasibility of the content-based image retrieval technology in educational resource platform.Secondly,with the above background and problems,this paper proposes a multi-feature image retrieval strategy based on Hadoop platform.it discusses the theoretical basis and methods of realization,gives a detailed analyzes of the key processes such as image feature extraction,similarity computing and parallel processing.Finally,under the guidance of the theory,I builded a prototype system and tested on the images of junior high school biology and Corel test set.The experimental data show that the multi-features retrieval strategy based on Hadoop platform can solve the problems of image retrieval in the current teaching resource library.At the same time,the result of retrieval is improved by comparing with the one computer environment,and it is more suitable for large-scale image retrieval.