The Design And Implementation Of Attractions Image Retrieval System Based On Android Platf Orm

With the development of tourism, the existing equipment of guide cannot follow the needs of tourists. In order to enhance the level of information of tourism, a new type of software is designed for guiding which is more convenient. It uses a variety of image features and hash algorithm to achieve image retrieval. The combination of Image Retrieval Technique and Android Platform makes a new guide system. The system can retrieve the image of the attractions; can make tourists get the information of attractions easily and improve the quality of tourism.By analyzing and comparing the status of research of existing guide system, an Android-based guide system is designed to help tourists to determine the travel destinations with the pictures but the name of attractions. This paper introduces the development mode of Android application and the technology of image feature extraction. Image retrieval technology is the key technology of this system.In the overall design of the system, the system is divided into two modules: One is Android-based client module, and the other is server image retrieval module. Client module can take photos of sights, check local images, upload images and check the result of retrieval. Sever module can deal with the images in background, receive the images from clients, do image retrieval and send the result of retrieval back to clients.This paper does the research of texture features, RGB color features and Gist feature of images. And combine these three features and Hash algorithm to finish the function of image retrieval. During the testing and implementation of the system, 2000 open source images are used. First, according to the image texture features classification relevance of the original data set is divided into four subsets, after a lot of test data obtained can be divided into subsets of boundary value. After evaluation of all the corresponding image correlation values obtained approximate range of values, should ultimately determine the value of 3 to divide subset. Then use the RGB color normalization feature of the above four subsets segments, similar to the final classification of the most Sights feature image according to Gist.This paper achieved the initial target, which is developing an Android-based guide system. Tourists can do image retrieval to get the related information of sights. It’s important to help the tourists to make sure where they are and make their travel plan.