4.2.4 Point Cloud Analysis
4.2.4 Point Cloud Classification using the Multiscale Curvature Algorithm
The Multiscale Curvature Classification (MCC) algorithm was developed especially for forest environments by Evans and Hudak in 2007. The two scientists created this algorithm that automatically and objectively classifies LiDAR data with only two classification parameters and minimized post-processing requirements. The MCC algorithm is used especially in forest conditions because it minimizes the commission errors and retains a big amount of ground points providing better quality of Digital Terrain Models. Classifying points as ground or non ground under forest conditions is a challenging process because of the ground convolution and the vegetation returns. The usual …show more content…
The point cloud is very dense (the lidar post spacing is 0.3 m or approximately 10 points are included in 1 m2) covering all local surface variations. Therefore the IDW interpolation technique is proper and can be used as it evaluates the influence of each point to the unknown value depending on their distance (Azizi and Nazafi, 2014). In other words IDW interpolation controls the distance to which cell samples affect the surface derivation. By taking into account the nearest points to the unknown location, the interpolated surface has more details. Furthermore, IDW allows to determine the power value of it, which specifies to which extent the result is localized or averaged. The higher this power value is, the more localized is the result, because the influence of the surrounding points decreases more rapidly with the distance. In this way depressions and peaks are well captured, fact that is important in this study. Power value of 2 gives good results (Qinghua et al., 2010) therefore it is selected here. The peaks and the valleys are not flattened because the highest and lowest z value of the features in the study area, are included in the dataset, since the point cloud covers meticulously the whole study area. The IDW is applied based on a kernel radius of 1m. Hillshade is used for better DTM visualization and we can detect easier the DTM differences based on different point cloud classification
The Multiscale Curvature Classification (MCC) algorithm was developed especially for forest environments by Evans and Hudak in 2007. The two scientists created this algorithm that automatically and objectively classifies LiDAR data with only two classification parameters and minimized post-processing requirements. The MCC algorithm is used especially in forest conditions because it minimizes the commission errors and retains a big amount of ground points providing better quality of Digital Terrain Models. Classifying points as ground or non ground under forest conditions is a challenging process because of the ground convolution and the vegetation returns. The usual …show more content…
The point cloud is very dense (the lidar post spacing is 0.3 m or approximately 10 points are included in 1 m2) covering all local surface variations. Therefore the IDW interpolation technique is proper and can be used as it evaluates the influence of each point to the unknown value depending on their distance (Azizi and Nazafi, 2014). In other words IDW interpolation controls the distance to which cell samples affect the surface derivation. By taking into account the nearest points to the unknown location, the interpolated surface has more details. Furthermore, IDW allows to determine the power value of it, which specifies to which extent the result is localized or averaged. The higher this power value is, the more localized is the result, because the influence of the surrounding points decreases more rapidly with the distance. In this way depressions and peaks are well captured, fact that is important in this study. Power value of 2 gives good results (Qinghua et al., 2010) therefore it is selected here. The peaks and the valleys are not flattened because the highest and lowest z value of the features in the study area, are included in the dataset, since the point cloud covers meticulously the whole study area. The IDW is applied based on a kernel radius of 1m. Hillshade is used for better DTM visualization and we can detect easier the DTM differences based on different point cloud classification