A 3D Reconstruction Method for Group Target Split Detection

The group target in 3D space is composed of a large number of individuals. In the process of group target movement, some individuals that can form a small group may be split from the whole. To achieve stable tracking of group targets, it is necessary to detect the split of group targets in time. Therefore, this paper proposes a 3D model reconstruction method based on 2D information to detect the split of the group targets. It uses the reconstructed ellipsoid to represent the shape of the 3D space group target and then the split judgment can be made by the volume change rate of the ellipsoid. Due to the calculation process of directly obtaining the ellipsoid parameters in the 3D space is complicated and the calculation amount is large, this method uses the projection idea to project the ellipsoid onto the coordinate plane, and then the projected ellipse parameters can be used to reconstruct the 3D ellipsoid. Simulation results prove the effectiveness of the new method.