The traditional method of scene recognition mainly focuses on the description of image features. Therefore, in this paper, we will discuss a method to mine the positioning information of SoOP, and use that information in scene recognition. There is insufficient usage of the relationship between positioning information and the scenery semantic, for example, when users are in an indoor environment holding their mobile devices, through the positioning signals, we can obtain the positioning information that relates to the scenery’s geometry structure then, we can obtain the semantic information of the current position by combining the positioning information with the elements that describe the scene (such as the indoor map). However, current research is mainly focused on the aspects of implementation or auxiliary positioning.
This indicates that SoOP can work in combination with other positioning sources to provide accurate, reliable, and real-time positioning. achieved positioning accuracy of up to 2–5 m by using the fusion signal of the built-in sensor, Wi-Fi, and magnetic field.
#INCEPTION WE NEED TO GO DEEPER SCENE BLUETOOTH#
In indoor positioning, which uses the mobile terminal as the development platform, Signals of Opportunity (SoOP), such as Wi-Fi or Bluetooth signals, are not specifically built to be positioning signals, but these signals have become a widely-researched topic of indoor positioning, because they are easy to build up and are cost-effective. Mobile terminals can provide basic communication information and perception for positioning and scene recognition technology. Since the mobile terminal contains built-in physical devices, such as vision and multi-sensors, it has become a new type of scene perception and communication platform for image acquisition and decision-making. The acquisition of real-time location information on mobile devices has become an indispensable element of intelligent devices. At the same time, with the improvement of hardware and developments in behavior calculation, more and more intelligent devices are being used in scientific research and real-world scenery. Innovative positioning technologies and wireless networks promote the development of the indoor positioning system. The accuracy of indoor scene recognition is improved in particular, the error rate at the spatial connection of different scenes is decreased, and the recognition rate of similar scenes is increased. The experimental result shows that, for two different experiment sceneries, the accuracies of the prediction results are 97.0% and 96.6% using the proposed model, compared to 69.0% and 81.2% by the method of overlapping positioning information and the base map, and compared to 73.3% and 77.7% by using the fine-tuned Inception V3 model. Using the method of transfer learning on the Inception V3 network model feature information is added to assist in scene recognition. Based on this, we design a convolutional neural network (CNN) model that concatenates features of image data and signals of opportunity for localization by using indoor scene datasets and simulating the situation of indoor location probability. The mobile device cannot only capture images of the indoor environment in real-time, but can also obtain one or more different types of signals of opportunity as well.
With the rapid development of indoor localization in recent years signals of opportunity have become a reliable and convenient source for indoor localization.
0 kommentar(er)
