Research And Development Of Parallel Terrain Factor Extraction Algorithm Based On Large-scale DEM

Natural disasters and over-exploitation aggravate soil erosion,soil erosion and other problems.Soil erosion monitoring can quantify soil erosion.Generally,the soil erosion equation(USLE)or the Chinese soil erosion equation(CSLE)can be used to calculate the soil erosion and soil erosion in a certain area.Slope factor(S)and slope length factor(L)are two important parameters in the soil erosion equation,collectively called terrain factor(LS),which can be used to measure the terrain of the region and reflect the impact of terrain on soil erosion.It can be seen that obtaining topographic factors is an important work for calculating soil erosion.Digital Elevation Model(DEM)is a kind of solid terrain model used to represent the elevation of the ground.It often exists in the form of raster images.The terrain factor extraction method based on DEM has gradually formed and widely used.With the more accurate and convenient acquisition of DEM,the accuracy continues to improve and the data scale continues to expand.However,the storage space and calculation amount of terrain factor calculation based on large-scale and high-resolution DEM are very large,and the efficiency of single-node calculation is low,so this paper intends to study the parallel extraction algorithm of terrain factor based on large-scale DEM.The main research contents of this paper include:firstly,the Priority-Flood parallel filling algorithm is optimized.For the case that a large scale DEM is input and not divided,optimization methods such as the segmentation strategy with halo are designed.The operation of the original algorithm to build the global minimum overflow elevation is changed from synchronous to asynchronous,which saves the time waiting for the construction of adjacent tiles.In the 260 billion pixel filling experiment,the optimized algorithm is consistent with the original algorithm,and the processing time is 37%shorter than the original algorithm,which improves the computational efficiency of parallel filling.Secondly,based on the pre-processing results of filling depression,the idea of three-step parallel framework is applied to the slope extraction,flow direction,concentration accumulation and other processing,and the parallel design based on Spark technology is carried out to realize the parallel extraction of large-scale DEM terrain factors,which greatly improves the computing performance and achieves the extraction results within the allowable range of error.Finally,the above parallel extraction algorithm is integrated into the dynamic monitoring platform of soil and water conservation as the terrain factor calculation module and the basic operator of soil loss equation to realize the dynamic monitoring of soil erosion and soil erosion in the region.This topic explores the algorithm of soil erosion terrain factor and applies big data tools and solutions to the field of geography.It can solve the practical problems in the calculation of water and soil loss,and also promote the application research of computer in geography.