Augmented reality navigation method based on indoor natural scene image deep learning

Abstract

The invention discloses an augmented reality navigation method based on indoor natural scene image deep learning. The method comprises the steps of firstly scanning an indoor natural scene through a three-dimensional laser scanner to extract three-dimensional scene feature recognition points, then calculating an internal reference matrix of a smart phone camera, collecting an indoor natural sceneimage through a smart phone to extract two-dimensional image feature recognition points, and establishing an indoor natural scene topological network structure chart through an indoor planar map; binding and mapping the two-dimensional image feature recognition points, the three-dimensional scene feature recognition points and the topological network path nodes through specific descriptors; carrying out deep learning-based image classification on the indoor natural scene images acquired by the smart phone, and segmenting the indoor natural scene into a plurality of sub-scenes; and then tracking and recovering the three-dimensional scene feature recognition points by using an optical flow tracking algorithm, further synthesizing a three-dimensional registration matrix required by scene registration, and finally finishing virtual navigation object real scene registration and finishing indoor natural scene path navigation.

Description

Technical field: [0001] The invention relates to augmented reality and indoor navigation technology, and belongs to the field of combining computer vision and augmented reality, in particular to an augmented reality navigation method based on deep learning of indoor natural scene images. Background technique: [0002] Augmented reality (Augment Reality, referred to as AR) is a hot spot in scientific research in recent years, and has a wide range of application prospects. By combining virtual objects with the real environment, people's cognition of the surrounding environment can be enhanced. The characteristics of augmented reality are: combination of virtual and real, real-time interaction and tracking registration. It superimposes computer-generated information (images, models, animations, etc.) on real scenes to achieve seamless integration of virtual and reality; augmented reality is different from VR's fully virtualized real world, but a supplement to the real environme...

AI Technical Summary

Problems solved by technology

However, deep learning of natural scene images and indoor augmented reality navigation on smartphones is still a difficult and open problem.

[0005] The above-mentioned various indoor positioning and navigation technologies either locate the local position through various signal inductions, or need to arrange artificial markers, and there are problems such as unstable signal sources, low positioning accuracy, and high failure rate; at the same time, such as indoor The method of placing a specific QR code has the problems of low real-time performance and poor interactivity. On the other hand, markers such as QR codes are easily damaged by human beings, which increases the difficulty of registration and navigation.

In view of the above factors, it is difficult to popularize indoor augmented reality navigation based on natural scene images.

Images