Research On Scan Line Filtering Algorithm And Its Application In Airborne LiDAR Point Cloud

LiDAR measurement technology is a brand-new technical means to obtain geospatial information with high temporal and spatial resolution,which enables us to change from traditional manual single-point data acquisition to continuous automatic data acquisition.It is also widely used in vegetation protection,three-dimensional terrain modeling and other fields because of its high-precision and intelligent characteristics.Due to the complexity of the terrain,point cloud filtering has always been a difficult problem in post-processing.This paper categorizes the existing point cloud filtering algorithms,and specifically introduces the development history of the popular algorithms,especially the scan line filtering algorithm that this article focuses on,and explains the basic principles,ideas,advantages and disadvantages of algorithms.There has never been a unified standard for the threshold setting of the filtering algorithm.This article attempts to obtain a unified standard for the threshold setting to realize the adaptive filtering of the scan line filtering algorithm.Through the method of statistical research,the histogram statistics of the various data characteristic indicators of the scan line filtering algorithm are carried out,and the best thresholds of the scan lines are obtained through experiments,and the changes in the size of the best thresholds and the histogram are obtained.The changes of each data feature are compared,and the data feature related to the optimal threshold is inferred.Finally,through Pearson correlation analysis,the correlation between the inferred data features and the optimal threshold is further proved.The correlation law between the data feature and the optimal threshold can provide a reliable basis for the pre-setting threshold of the scan line algorithm.This paper proposes the concept of vertex vector angle in view of the insufficient terrain applicability of the scan line filtering algorithm,especially the error-prone problem of judging the noise points below the ground and the points with large differences between the positive and negative slope angles.The vertex vector angle is added as another judgment condition to the traditional scan line algorithm,and the obtained vertex vector angle-scan line filtering algorithm can filter out the above two types of non-ground points very well.The definition of the vertex vector angle and the flow of the vertex vector angle scan line filtering algorithm are introduced in detail,and three sets of experiments prove that the vertex vector angle scan line filtering algorithm proposed in this paper can improve the terrain applicability of the scan line filtering algorithm.This article analyzes the future development direction of such algorithms by starting from the core principles of scan line filtering algorithms.Through the specific introduction of manual editing,the similarities between the scanning line filtering algorithm and manual editing are found.The algorithm development direction of scanning line filtering algorithm instead of manual editing is proposed.Through experiments,the scanning line filtering algorithm is used as a post-processing and clustering filtering algorithm for combined filtering,and the topographic map formed by the ground points obtained before and after the combination is compared.And the comparison of filtering accuracy verifies that the post-processing of the scan line filtering algorithm improves the filtering accuracy of the original filtering algorithm,and also verifies the possibility of the scan line filtering algorithm to replace manual editing work.It is further proposed that manual editing is not only the application development direction of scan line filtering algorithm,but also the development direction of scan line filtering algorithm optimization research.