Research On Indoor Localization Based On PROSAC Algorithm

With the development of society's intellectualization,more and more intelligent mobile devices have become popular in our daily life.Therefore,people are increasingly demanding Location Based Services.In the Location Based Service system,the acquisition of location information is the most important core part of the system,and how to accurately obtain the user's location information in the scene has also become a hot issue of research.In outdoor environments,satellite signals such as GPS can offer accurate positioning to users.However,due to the shielding effect of buildings and other factors,it is difficult to obtain GPS signals in indoor environments.Therefore,in a complex indoor environment,researchers have proposed a computer vision-based indoor positioning method for this problem.The visual indoor positioning method has become a promising indoor positioning method due to the convenience of visual information acquisition,low equipment deployment cost,and rich feature information.This paper first introduces the purpose of research on the visual indoor positioning system,the research status of domestic and foreign indoor positioning technology,and the current research status of visual positioning technology.Secondly,the relevant basic theories in the visual positioning system are studied.Finally,based on this,the main contents of this paper are studied for:(1)For the problem that the image retrieval process in the existing visual positioning system takes a long time,this paper proposes a hybrid retrieval algorithm based on Wi Fi and images.By introducing Wi Fi data source as an auxiliary source for image database retrieval,we use the rough match of Wi Fi signal strength to select the area of the area where the user query image is located.Then,the searching scope of the image database is reduced,and the method of fine matching of image features is used to search the similar images in the database.The method which combines the Wi Fi rough-matching with the image fine-matching can improve the traditional image retrieval process and reduce the computational complexity of the traditional image retrieval process.At the same time,the image can be more quickly and efficiently performed under the premise of ensuring the accuracy of image retrieval.(2)The PROSAC algorithm is proposed to solve the problems of unstable RANSAC algorithm.It can reduce the large amount of calculation and large time overhead of the RANSAC algorithm applied in the process of eliminating mismatched points in the process of traditional image matching.Due to the fact that the sampling points are randomly selected when the RANSAC algorithm estimates the model,the sampling points which are randomly selected may not be all internal points.Therefore,the sampling points should be reselected to continue estimating the model.This results in the unstable iteration times,large amount of calculation and long time-consuming shortcomings of the RANSAC algorithm.In the PROSAC algorithm proposed in this paper,the quality factor ? is introduced to sort the qualities of the sampling points.The high-quality internal points are preferentially selected to estimate the model,thereby reducing the number of iterations and improving the stability of the algorithm.