Method and device for acquiring three-dimensional scene image by using fisheye camera, equipment and medium

Abstract

The invention provides a method and device for acquiring a three-dimensional scene image by using a fisheye camera, equipment and a medium. The method comprises the following steps: acquiring a video image of a target scene by using the fisheye camera; and fitting the video image into a three-dimensional space model based on a corresponding relationship between the three-dimensional coordinates of the vertexes in the three-dimensional space model of the target scene and the corresponding two-dimensional coordinates to obtain a three-dimensional scene effect picture of the target scene, wherein the three-dimensional coordinates of the vertexes are spatial position coordinates of the vertexes in a three-dimensional coordinate system corresponding to the three-dimensional space model, and the two-dimensional coordinates are coordinates of projection points of mapping points corresponding to the vertexes on a cross section of a spherical coordinate system corresponding to the fisheye camera; the spherical coordinate system is constructed by taking the fisheye camera as a sphere center based on internal parameters and external parameters of the fisheye camera; the cross section is a plane which passes through the central point of the spherical coordinate system and is vertical to the positive shooting direction of the fisheye camera. The image acquisition cost can be reduced, the imaging effect of the 3D scene is improved, and the real-time requirement of the video stream is met.

Description

technical field [0001] The present disclosure relates to video image processing technology, in particular to a method, device, device and medium for acquiring a three-dimensional scene video image by using a fisheye camera. Background technique [0002] At present, the imaging methods in applications such as augmented reality technology (Augmented Reality, AR) all collect the model data and material data of the entire scene through data modeling, and use these data models and material data through a certain three-dimensional (3D ) technology to restore the entire 3D scene. [0003] In the process of realizing the present disclosure, the inventors have found through research that the above-mentioned imaging methods in applications such as AR have at least the following problems: the model data and material data of all objects in the entire scene space need to be collected in advance for modeling, and cannot Use the video stream to restore the real imaging of the entire 3D sc...

AI Technical Summary

Problems solved by technology

[0003] In the process of realizing the present disclosure, the inventors have found through research that the above-mentioned imaging methods in applications such as AR have at least the following problems: the model data and material data of all objects in the entire scene space need to be collected in advance for modeling, and cannot Use the video stream to restore the real imaging of the entire 3D scene in real time; at present, most of the ordinary cameras are used for image acquisition. Due to the limitation of the shooting angle of ordinary cameras, the imaging of the entire space is usually achieved by fusing the images taken by multiple cameras. The image acquisition cost required to deploy multiple cameras is high, and if there are walking objects or humans in the space, there is a high probability of splitting at the fusion gap of the images captured by multiple cameras, thus affecting the imaging effect of the 3D scene

Images