Research on Depth Map Enhancement Method Based on Bilateral Grid
Depth information plays an irreplaceable role in many computer vision tasks,but it is difficult to obtain accurate depth information.At present,there are some problems in the way of obtaining depth information,such as low resolution of depth map,wrong depth value and so on.This paper proposes a new optimization model to solve the above problems.Different from the ex-isting model that only applies confidence on data term,the model also introduces complementary confidence on smoothness term,which can well alleviate the artifact problem caused by excessive weight of smoothness term,and improve the performance of the model.In addition,this paper solves the above optimization model in the bilateral grid,which greatly reduces the running time.The experimental results show that the proposed method can enhance the quality of depth map,make it meet the requirements of rele-vant applications,and has high efficiency.
国家科技期刊平台
NSTL
万方数据
