Research On Smart Space Positioning Technology Based On Projection Method

In modern society,the life demand for convenience and safety is getting higher and higher.Especially,elderly people who are not able to move are more likely to need intimate life care.With the increasing trend of population aging,the research of family service robot is becoming more and more extensive.The organic combination of the smart space and the service robot is a research hotspot nowadays.A service robot is mainly responsible for identifying,navigating,locating and crawling,etc.The smart space senses the environment through various sensors and detection devices.It can provide more environmental information for the service robot and assists the service robot in performing various tasks better.Aiming at the problem of the existing positioning and ranging methods depended on the external conditions such as the number of sensors,artificial markers and the ground,this paper proposes a positioning method based on fringe projection.The specific contents are as follows:Firstly,a geometric imaging model is established and Zhang Zhengyou calibration algorithm is studied.The black and white checkerboard is used as the calibration plate pattern.The images of the calibration board of the projector and the camera are acquired at the same time.The intrinsic and extrinsic parameters are solved according to the world coordinates and pixel coordinates of the feature point.The system calibration is achieved and the spatial position relationship of the projector and the camera is obtained.Secondly,a ranging method combined with fringe projection and three-dimensional reconstruction is proposed.The width of the fringes is calculated by ZS(Zhang-Suen)thinning algorithm.The relationship between the projection distance and the width of the fringes is determined according to the projection model.Then the distance between the reference surface and the projector is determined.The wrapped phase map is obtained by the four-step phase-shift method and the unwrapped phase map is obtained by the least-squares phase unwrapping method.The height of the measured object and the three-dimensional topography are determined according to the grating projection principle.Thus,the distance between the object and the projector is determined.The mean square error of the experimental ranging result is 0.05 m and the feasibility of this method is verified.Finally,in order to obtain large-field images of the indoor space,this paper also studied the image stitching technology.For the grating fringe images,the traditional image stitching method based on Harris is improved.The corner features extracted by the Harris algorithm are combined with edge features of the fringe extracted by the Canny algorithm to perform feature matching and stitching.Experimental results show that the improved stitching algorithm based on grating fringe images can improve the accuracy of feature matching by more than 20%.