Unmanned aerial vehicle real-time path planning method for urban scene reconstruction

Abstract

The invention provides an unmanned aerial vehicle real-time path planning method for urban scene reconstruction, which uses a scene top view as priori information to generate an initial path of an unmanned aerial vehicle, optimizes the initial path in real time in the flight process, and realizes three-dimensional reconstruction of an urban scene. The method mainly comprises the following four steps: step 1, a scene top view is analyzed to obtain a scene layout, and an initial path of the unmanned aerial vehicle is generated; step 2, a sparse point cloud of a building is reconstructed in the flight process according to the initial path, the height of the building is estimated, a rough scene model is generated in combination with the scene layout, and the height of the initial path is adjusted; step 3, by means of the rough scene model, the sparse point cloud and the flight path of the unmanned aerial vehicle, a scene coverage confidence map and detail parts required to be close-up areobtained, and the flight path is optimized in real time; and step 4, a high-resolution image is obtained, and a three-dimensional model of the scene is obtained by reconstruction. The method has feasibility, accuracy and universality and can be applied to three-dimensional reconstruction, automatic surveying and mapping and the like of cities.

Description

technical field [0001] The invention relates to a real-time path planning method for UAVs used for urban scene reconstruction. The UAV initial flight trajectory is generated by using the top view of the scene as prior information, and the initial trajectory is optimized in real time during the flight process. The image acquisition and three-dimensional modeling of the scene have certain effectiveness and versatility, and belong to the field of robot path planning and computer graphics. Background technique [0002] With the rise of "digital cities", the 3D reconstruction of urban scenes is becoming a current research hotspot. The construction of 3D models of urban scenes can be widely used in urban planning, urban supervision, urban image promotion, real estate development and other fields. However, because the urban scene environment is relatively complex and the scale is relatively large, 3D reconstruction is also quite difficult. Because drones can collect data in the ai...

AI Technical Summary

Problems solved by technology

The invention has strong robustness, solves the time-consuming and labor-intensive problem of using a rough model as a priori for path planning in the prior art, and overcomes the limitation that the current technology is difficult to generate a three-dimensional model of a city with details

Images