Abstract:

In recent years,visual positioning is widely used in indoor positioning and navigation because of its high positioning accuracy.However,we can't achieve high-precision positioning in many scenes with visual positioning algorithms due to complex indoor environment,and most of the visual positioning algorithms are not suitable for smartphones for its' high time and computational complexity.In order to realize the accurate and efficient positioning with smartphones,a method based on similarity computation of gray histogram is proposed.The method is divided into two stages:the offline stage and the online stage.In the offline stage,we acquire the image in the grid points of the known coordinates with the smartphone,and then the gray histogram of the image is extracted to establish the image gray histogram image feature database;In the online stage,the image is collected and gray histogram is extracted on test point by the smartphone,and then calculate the similarity between it and the image feature database.The grid of similarity and maximum is selected as the approximate position,the similarity is used as the weight,and the exact location is obtained by weighted mean method.Compare the proposed method with the KNN algorithm,the experimental result shows that the average positioning error is less than 0.3 m,average point positioning accuracy of this method is better than that of KNN algorithm,and compared to the KNN algorithm,the positioning accuracy is improved by 40.7%,and the computing time is better than 1.7 s.

Key words: indoor positioning, gray histogram, image feature database, similarity calculation, weighted mean