Method for acquiring position and posture of images within city range based on deep learning

Abstract

The invention provides a method for acquiring positions and postures of images within a city range based on deep learning, relating to the field of image geographic position recognition and augment reality. The method comprises the following steps: 1) creating a city picture set; 2) training a Gaussian mixture model to the city picture set, and dividing city geographical areas with the trained Gaussian mixture model; 3) training a combined learning picture posture estimation and scene recognition nerve network; 4) initializing, uploading the GPS or approximate network position information of auser; 5) classifying the approximate position information by using a learning partition function, downloading a corresponding network model and rendering information needing displaying to a user side; and 6) collecting a user input camera video steam, predicting the location results of three levels at the current moment by applying the downloaded network model of the current area, and if the confidence coefficient of the prediction result output by the network is higher than the threshold value, performing rendering of the rendering information by using the predicted position and posture parameters.

Description

technical field [0001] The invention relates to the field of image geographic location recognition and augmented reality, in particular to a method for obtaining the location and posture of an image within a city range based on deep learning. Background technique [0002] With the explosive development of mobile Internet and smart devices, taking and sharing photos has become a part of people's daily life. How to deduce from the photo where the photo was taken and the perspective of the photo has become a very meaningful problem. The problem of inferring the shooting position and perspective from the photo is also called the pose estimation problem of the camera in Multi-View Stereo. It is a basic problem in the field of computer vision and robotics, and has a wide range of applications, such as enhancing Reality (Augmented Reality, referred to as AR), simultaneous positioning and map construction (Simultaneous Localization and Mapping, referred to as SLAM), and image-based...

AI Technical Summary

Problems solved by technology

Although the accuracy of coordinate regression forest is very high, its disadvantage is mainly that RGB-D images are required as input. In actual use, RGB-D images are only suitable for indoor scenes, and the RANSAC calculation process is very time-consuming.

Images