Abstract: The present disclosure provides a method and a device for interactive calibration based on 3D reconstruction in a 3D surveillance system. The method includes: acquiring a reference background model and a surveillance video, capturing multiple scene images, and connecting multiple surveillance cameras corresponding to the surveillance video with each other, wherein at least one scene image of the multiple scene images has an overlapping portion with a site of the surveillance video; performing 3D reconstruction based on the multiple scene images to generate a 3D feature point cloud of the scene, and embedding the 3D feature point cloud into the reference background model; estimating an attitude of a current frame of the surveillance video, and automatically calibrating the surveillance camera; and calculating a homographic transformation from the attitude of the current frame to the reference background model, and embedding an image projection into the reference background model.

Abstract: The present disclosure provides a method and a device for interactive calibration based on 3D reconstruction in a 3D surveillance system. The method includes: acquiring a reference background model and a surveillance video, capturing multiple scene images, and connecting multiple surveillance cameras corresponding to the surveillance video with each other, wherein at least one scene image of the multiple scene images has an overlapping portion with a site of the surveillance video; performing 3D reconstruction based on the multiple scene images to generate a 3D feature point cloud of the scene, and embedding the 3D feature point cloud into the reference background model; estimating an attitude of a current frame of the surveillance video, and automatically calibrating the surveillance camera; and calculating a homographic transformation from the attitude of the current frame to the reference background model, and embedding an image projection into the reference background model.