Research Of Smart Phone Indoor Positioning Accuracy Evaluation System Based On Encoded Targets

Indoor positioning is one of the key technologies for future artificial intelligence and super intelligent applications.With the popularity and development of smartphones,indoor positioning technology based on smartphones has a broad market application prospect and an important social significance.At present,there are some mature indoor positioning methods including RF signal-based positioning technology,sensor-based positioning technology,and multi-source fusion positioning method.At present,the evaluation methods for the positioning accuracy of these indoor positioning technologies generally use higher-precision positioning instruments such as measuring robots and motion capture systems.However,these approaches are difficult to carry out the evaluation directly and dynamically in real time,which makes the final evaluation results have certain errors.It is difficult to obtain the ground truth in real time for indoor positioning using smartphones directly,dynamically,and precisely.To solve this problem,this paper proposes and implements a real-time evaluation system for smartphone positioning using color visual scatter-encoded targets based on machine vision and photogrammetry.It is a robust and dynamic solution with high precision.It can accurately and quickly obtain the dynamic instantaneous pose of a smartphone camera and offer a precise realtime ground truth via a visual positioning technology based on artificial scatter feature.It is an accurate,fast,efficient and low-cost accuracy-evaluation solution for smartphone positioning indoors.The main researches and innovation points of this thesis are summarized as follows:1.A set of novel high-precision color vision scatter-encoded patterns with a robust recognition rate is designed.The encoding method of the scatter-encoded targets designed in this thesis combines the color information and geometric information.It is solution with a stable,high recognition rate and sufficient encoding capacity.2.The dataset of the encoded targets is built based on color vision scatterencoded patterns.We use a smartphone to obtain a sequence of images from an experimental room and extract the base points of the color visual scatter-encoded patterns in order to establish the local coordinate system of the encoded targets.According to the coordinates of four control points in the local coordinate system,the coordinates of all object points can then be transformed to the local coordinate system.3.We proposed a real-time evaluation system based on encoded targets for smartphone positioning indoors.We use an efficient algorithm to decode the targets from the image taken dynamically in real-time to obtain a precise instantaneous pose information of the smartphone camera for the epoch of taking the image.With this approach,we can establish the ground truth can be used for all smartphone positioning technologies.