Dokładność szybkich metod inwentaryzacji leśnej mapy numerycznej

The functioning of the forest digital maps and conducting complex silviculture implemented in accordance with the diversified development creates great demand for a quick method of spatial data capture. Currently used data collectors and navigation receivers enable the measurements of the forest, but the results are not reliable taking into account their quality. In this study the evaluation of currently used measuring devices were realized and new, reliable solution that can streamline data capture in the field was proposed. The work was carried out in the Pułtusk Forest District (central Poland). We established five plots in the stand with the alder as a dominant species. 137 reference points were captured using classical surveying methods. We used the following data collectors: Motorola Omnii XT15 and Workabout PRO, Samsung Galaxy Core Plus smartphone and GIS Trimble ProXH navigation receiver. In addition, the unmanned aircraft i.e. popular on the market DJI Phantom 4 was also used. The accuracy was assessed by determining the error of the plot area and the accuracy of control points (MP). It was found that current popular data collectors and smartphones define the plot area with an error of about 3%, the same positioning of the object can be accurate from 2 to 6 meters (tab. 1). The results allow to conclude that the use of smartphones may be important only for draft assessment, and the realization of the measurement using the data collectors should always be preceded by the control measurement on the charac− teristic points. Significantly better results are obtained by the GIS class receivers (MP less than 2 meters) and using differential correction (MP less than 0.6 meters). However, these devices are used rarely in practice, despite universal access to the network of GNSS reference stations. The use of drone allowed to obtain images for the entire area in 15 minutes. The orthophotomap was automatically made in point−cloud technology. This method gave an error of 1.3% for estimat− ing the plot area and 1.69 meters for the position of control points for objects borders. Given the falling prices of unmanned platforms and low labor intensity of this measurement method, we can assume that it will find widespread use in updating of forest digital maps.