| |
Abstract |
When describing a LiDAR dataset, many aspects are unambiguous such as the area of coverage or the acquisition date. Additional characteristic values of accuracy and error typically accompany the data and well-defined guidelines exist for how these values should be derived and reported. Two supplementary characterizations are frequently used, namely Nominal Spacing and Density, for which there is no standardized method on how they should be derived and reported. A statistical bias is easily introduced when providing spacing and density quantification. A method of measurement is presented in which spacing and density statistics can be qualified and bias identified. The law of large numbers certainly applies to datasets with millions or billions of points and means that the variance can be reduced, but not the bias. While it may appear trivial, the principal contributors to LiDAR spacing and density bias are the absence of clear and concise definitions. Bias can not be entirely eliminated but should be reduced wherever possible. |
|
