Research On Method Of 3D Video Reconstruction Based On 2D Video

3D video has been the focus of attention. Because of the high cost and long period of 3D video production, 3D video sources can't meet the needs of the market development. It has become one of the research directions of the new 3D digital video that how to automatically convert the two-dimensional video to three-dimensional video with a real sense.Depth estimation is one of the key technologies for converting 2D to 3D. At present, the majority of 2D video is composed of static background and moving objects. According to the characteristics, a hybrid depth map extraction algorithm is proposed, which uses the foreground and background respectively to obtain depth information. Firstly, the static background is extracted based on the gaussian mixture model and foreground objects are obtained by the background subtraction division. Then, the background depth map is obtained based on linear perspective according to the rigid geometry line feature and the prospect depth map is extracted based on motion information. In order to better show the depth details, laws template is selected to obtain the depth values of texture properties. Finally, the hybrid depth map is generated by weighted fusion. Experimental results show that the proposed method can obtain the depth map effectively and can better show the depth details of the image. The quality of the depth map is improved.Virtual view synthesis is another one of the key technologies for converting 2D to 3D. The red component of original image is left offset to acquire the left view according to the estimated disparity from the extracted depth map. Then, the hole filling algorithm based on parallax information is used to fill the void of the left view. Finally, the blue green component of original image is used as the right view, synthesize the right and left view to obtain the stereoscopic 3D video image. A method of 3D video reconstruction based on 2D video is acquired. The experimental results verify the validity of the proposed method to construct the 3D video.