Research On Optimization Method Of Visual Path Planning Based On Multi-resolution DEM

Viewshed calculation based on DEM(Digital Elevation Model)is the primary content of terrain visibility analysis.The problem of path planning based on the viewshed is one of the crucial applications of terrain visibility analysis,which is of great significance in autonomous obstacle avoidance of robots,urban road planning,and electronic map navigation.With the rapid development of computer and remote sensing technology,DEM terrain data is increasing.How to process and analyze large-scale DEM data effectively and quickly is difficult in path planning on terrain.In summary,this thesis researches the viewshed calculation and path planning on high-precision DEM data.The main work is as follows:(1)A calculation method of the viewshed based on the reference plane is proposed.The viewshed calculation method is the basis of the viewshed-related applications.This thesis studies the viewshed algorithm in-depth and suggests applying visual point division calculation and designing a viewshed calculation based on the reference plane method.This method starts from the viewpoint and selects the visible points according to the rules to divide the target point area of each layer to calculate the visible region.Through experimental comparison and analysis with other traditional viewshed algorithms,the method proposed in this thesis can meet the needs of viewshed analysis in terms of efficiency and accuracy.(2)The shortest path method based on a multi-resolution raster DEM data model is proposed.As the amount of DEM terrain data continues to grow,there is a high degree of complexity in path search on grid DEM terrain.Searching for the optimal path on the high-resolution raster DEM is very time-consuming.Thus,this thesis proposes a multi-resolution terrain hierarchical model representing the same terrain and finding the coarse-grained shortest path on the low-resolution terrain data.Then,according to the mapping mode and rules of the terrain hierarchy model,the path nodes on the low-resolution terrain are mapped to the high-resolution terrain data.Then the local shortest paths between adjacent grids are solved in parallel on the high-resolution terrain data.Finally,the local path is connected to get the global way.This method reduces the range of path searches and effectively improves the efficiency of solving the optimal approach.Experiments show that the method proposed in this thesis dramatically improves the efficiency of path search on high-precision terrain data.The accuracy of the solution meets application requirements.(3)A method of least visible path planning based on a multi-resolution raster model is proposed.First,each viewpoint's views on the original DEM terrain data are calculated and expressed with a view matrix.The input for the least cost path uses a view matrix.According to the view matrix,this thesis constructs the optimal model of the hidden and the landscape path.It then uses the shortest path method of the multi-resolution data model proposed to solve it.Experiments show that the hidden and landscape path planning methods based on the multi-resolution raster DEM have apparent advantages inefficiency.