Vision Based Real Time Indoor Navigation with Augmented Reality

Project Title

Vision Based Real Time Indoor Navigation with Augmented Reality

Project Area of Specialization Augmented and Virtual RealityProject Summary

Global Positioning System (GPS) is the technique that continuously tracks the position of users and their environment through satellite. However, GPS works only in outdoor environments due to lack of satellite signals indoors thus, making it difficult to get the exact location of the users inside buildings. Therefore, localization in indoor environments depends on other techniques such as Wi-Fi routers, Bluetooth Beacons and RSSI fingerprinting. These technologies are expensive and do not give accurate location of the user as they depend on estimations mainly. Due to these reasons, effective indoor navigation in environments like shopping malls, hospitals, airports and universities is still a challenge. The proposed system focuses on exploiting the benefits of real time computer vision for providing a methodology for effective indoor navigation. This application scans the real environment through live camera stream of user’s mobile phone and matches it with a pre-collected sequential database of images containing feature points and coordinates. The location of the user is obtained through live camera scanning by recognizing the environment. While the user walks along the route the proposed system keeps recognizing the path and update the precise location. The proposed system uses Gyroscope and Accelerometer to detect the orientation of the user and instructs to follow the required path.

Project Objectives

• To provide a mechanism for navigating in indoor environment, where GPS doesn’t work.
• To implement visual and auditory instructions for better understanding of the path.
• To enhance the level of accuracy in recognizing indoor environments.

Project Implementation Method

The development of the android application uses the following programming languages, tools and SDKs.

• JAVA (backend)
• XML (frontend)
• ARCORE SDK (Used for mapping and recognizing the real world)
• Sceneform (To render a 3d model for better understanding)
• Microsoft Azure Spatial Anchors (Cloud to store feature points)
• Firebase (To store the information of anchors)

Hardware Requirements

• ARCORE supported device.
• Device with Auto-Focus Camera.

Software Requirements to run this application

• Android 8 + operating system
• ARCORE installed
• RAM 2GB and above

Benefits of the Project

• Easy navigation inside complex buildings
• No external help needed to find the destination.
• Useful for universities, shopping malls, airports etc.

Technical Details of Final Deliverable

Visual recognition solutions work especially well when sampling of the area in question is done from different angles through a camera, where a so-called visual point cloud or anchor with feature points is stored in the cloud database, which the system will later use to locate the relative position by comparing the actual view with the previously collected feature points. Microsoft Azure Spatial Anchors have been used to store the anchors along with the feature points in the cloud. These feature points will then be recognized in real world and display the visual instructions, which are 3d models placed in real world with Augmented Reality.

Final Deliverable of the Project Software SystemCore Industry ITOther Industries IT Core Technology Augmented & Virtual RealityOther Technologies OthersSustainable Development Goals Sustainable Cities and CommunitiesRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	64000
Microsoft Azure Account	Equipment	1	20000	20000
Nokia 6.1 plus (ARCore Supported Device)	Equipment	1	34000	34000
Miscellaneous	Miscellaneous	1	10000	10000