Research And Implementation Of Terrain Feature Recognition Method Based On Flowing Process

As the most important feature lines of the terrain skeleton,the valley line and the ridge line have important applications in automatic geomorphology synthesis,hydrological analysis,and geographic information system construction.Therefore,it is necessary to identify valley lines and ridgelines quickly and accurately from topographic maps.The topographic feature recognition method based on the flow process is based on the natural law that the surface water flow tends to flow from the high terrain to the low terrain,then simulating the flow process,and transferring the topographic features to the accumulated water volume data.Compared with other terrain feature recognition methods,the extraction results are more consistent with global features and real landscapes.However,this method has many problems such as large amount of calculation and local "parallel lines".The thesis proposes an improved algorithm for these problems.The improved algorithm uses the water accumulation method of four-direction flow simulation to effectively reduce the amount of calculation,and then uses image processing technology to extract topographic features from the water accumulation matrix.The extraction results are more refined and smooth,more in line with the real terrain structure,and there is no "parallel lines" problem.The main work completed in this thesis is summarized as follows:1.From the three aspects of algorithm principle,design and implementation,and recognition effect,the traditional methods based on flowing process are deeply studied,and the existing problems and the direction to be improved are analyzed.2.For the problem of large computational complexity of traditional algorithms,the water accumulation scheme using four-direction flow simulation reduces the number of repeated and iteration calculations compared with the traditional algorithm,and can realize the water accumulation quickly and effectively.Moreover,the scheme does not require each grid water flow to be led to the edge of the terrain,at the same time,reduces the calculation of the process of determining the smooth flow direction and greatly improves the overall efficiency of the algorithm.3.For the water volume matrix that have completed the water accumulation,the improved algorithm add image processing techniques such as filtering,blank filling,and thinning.Compared with the traditional method of directly extracting terrain feature from the feature points,the extraction results are smoother and finer,more in line with the real terrain structure,and there is no "parallel lines" problem.4.The traditional algorithm and the improved algorithm are compared and analyzed from the aspects of time efficiency and extraction results by experiments,which proves the accuracy and effectiveness of the new algorithm.5.The scanning sequence,water threshold and thinning algorithm which may affect the extraction result in the implementation of the improved algorithm are tested experimentally.The appropriate parameters are selected according to the experimental results,and the superiority of the improved algorithm is further verified.Based on the experimental results of the improved algorithm,the experimental results compared with the traditional algorithm,and the running time of the two algorithms,the conclusions of the improved algorithm are better than the traditional algorithm in terms of both time efficiency and extraction efficiency,which proves the improved algorithm is a fast and effective method for terrain feature recognition based on flowing process.