Research On Landmarks-based Binocular Vision Indoor Localization Algorithm

Because of the rapid development of Internet and wearable devices,location-based services are gaining widespread attention.As people are living for more than 70% of their time in the indoor environment,the indoor localization technology has become a hot research.With the popularity of smart terminals,such as mobile phones and cameras,the whole built-in sensor device makes vision-based localization systems gradually replacing those indoor localization system requiring special design.Firstly,this paper expounds the purpose and significance of binocular vision localization technology,and analyzes domestic and international research status of vision-based localization technology and binocular vision localization technology.Meanwhile,the application of binocular vision system model,image local feature and other related theoretical basis in vision-based localization system are studied.In addition,this paper completes the following research on some problems existing in the offline stage and the online stage of the traditional vision-based localization system:(1)A method of establishing an offline database based on landmark is proposed for the problem of excessive workload when building the traditional location fingerprint offline database.This method can effectively reduce the storage capacity of the database by using SURF feature points of landmark in the scene image instead of the global fingerprint information.Furthermore,through the homography,it can decrease the complexity and save time overhead;(2)Aiming at the problem of low positioning accuracy in traditional monocular vision localization algorithm,an improved binocular vision localization algorithm is proposed,which effectively reduces the time cost of landmarks' feature matching and improves the positioning accuracy.In addition,this algorithm can calculate the user steering angle,which enriches positioning navigation information.The simulation experiments about establishing the offline database based on landmark and the improved binocular vision localization algorithm is presented.The results show that the proposed method can reduce the storage capacity of the database effectively,reduce the time complexity and save the manpower cost.Moreover,the improved binocular vision localization algorithm could improve the localization accuracy,and obtains the steering of the online users as well.