The Depth-Image-Based Rendering (DIBR) has been widely used in the field of stereoscopic video application. With DIBR, the virtual view under new perspective can be generated from the scene in existing plane video and the corresponding depth map. However, DIBR has an inherent problem, i.e. there is dis-occlusion area in the virtual view under new perspective, which is mainly execute on originally blocked background area in original perspective is exposed in the new perspective. Therefore, filling the dis-occlusion area in new perspective rapidly and efficiently is the hotspot in the research of DIBR technology. In traditional DIBR technology, the restoration in texture direction is not good, hence the geometric distortion and poor stereoscopic sensation may be caused. This thesis mainly focuses on the improvement of traditional DIBR technology in terms of the restoration of dis-occlusion area. The major works are as follows: 1. Introduce two solutions of dis-occlusion, i.e. multi-directional extrapolation hole-filling (MDEHF) method and gradient-based deformable template hole-filling (GBDTHF) method. Then, a new method to fill the dis-occlusion area with texture and gradient is proposed. 2. To fill the dis-occlusion region quickly, the thesis initiates the method based on texture synthesis and the gradient influence on color change. This is based on the relationship between texture direction and dis-occlusion area. Making full use of the fact that most natural images have certain textures, the method gives a method to calculate the main texture direction and then narrows the search scope to the main direction. Thus reducing the time of searching for the optimal match and improving the efficiency of image restoration. Also it defines a priority, trying to make the greatest priority with more existing information and more structure information. This method makes the block with the greatest priority repaired firstly. As the size of window has a great impact on the result of restoration quality, a kind of scheme iv selecting template window size by self adaptation is offered to improve the quality of image restoration. A lot of experiments prove that the algorithm shortens the time of restoration and achieve a good restoration quality.