Scene reconstruction data acquisition method, device, computer equipment and storage medium

Abstract

The present application relates to a scene reconstruction data acquisition method, device, computer equipment and storage medium. The method includes: capturing a scene image; performing real-time environment perception on the scene image, detecting an object to be reconstructed in the scene image, and reconstructing three-dimensional representation data of the object to be reconstructed; based on the three-dimensional representation data of the object to be reconstructed and a preset distance threshold , determine the reconstruction target in the object to be reconstructed, generate an exploration path according to the preset exploration path planning algorithm, and explore the area to be reconstructed according to the exploration path; when the reconstruction target is reached, collect the reconstruction target according to the preset reconstruction path planning algorithm Rebuild data. This method does not require the input of scene prior geometric information, greatly shortens the reconstruction period, and can shorten the data collection time, thereby effectively improving the reconstruction data collection efficiency.

Description

technical field [0001] The present application relates to the technical field of computer vision, in particular to a scene reconstruction data acquisition method, device, computer equipment and storage medium. Background technique [0002] With the rapid development of drone technology, how to use drones to autonomously and quickly reconstruct large-scale urban scenes has become a research focus. [0003] In order to achieve a good reconstruction effect, the UAV needs to design the best data collection path, fully observe the building from different perspectives, collect sufficient reconstruction data, and achieve full coverage of the scene. However, most of the current reconstruction data acquisition methods rely on an additional flight and reconstruction of the UAV to obtain the geometric prior information of the scene, which greatly increases the reconstruction cycle and difficulty, and does not require the geometric prior information of the scene Information methods suc...

AI Technical Summary

Problems solved by technology

However, most of the current reconstruction data acquisition methods rely on an additional flight and reconstruction of the UAV to obtain the geometric prior information of the scene, which greatly increases the reconstruction cycle and difficulty, and does not require the geometric prior information of the scene Information methods such as DJI-Terra cannot be well adapted to different scenarios

More importantly, all algorithms calculate the path offline. When the pre-calculated scene information is inaccurate, the quality of the planned path will be reduced accordingly, making the collection time longer, and there may even be safety issues of collision with the scene.

[0004] It can be seen that the existing data collection methods for scene reconstruction have the problem of low data collection efficiency.

Images