Silvilaser 2019 - Poster Presentations »
Application of terrestrial laser scanner for the determination of forest inventory parameters using simple scans
The laser scanner technology (TLS) has been gaining strength as an alternative in acquiring accurate and detailed forest information, bringing improvements to conventional forest inventory techniques. Several studies have been carried out and show that one of the main bottlenecks in the use of simple scans is due to the partial detection of trees due to the shading effect caused by the interposition of other individuals, thus underestimating the number of trees and subsequently interfering with estimates in the plot. In this sense, the objective of the present research was to perform the automatic detection and determination of the diameter at breast height (DBH) of individuals of Pinus taeda L. by the TLS survey using simple scans, in order to verify if there are significant differences between the method TLS and conventional inventory. With densities ranging from 1111 to 2500 arv.ha-1, the population evaluated covers an area of 4.1 hectares with 4 spacings. The total was installed 11 circular sample units of 300 m² and were measured the DBH and total height all individuals. The TLS survey was performed with the Leica BLK 360 equipment, allocating the equipment at the central point of the plot. The laser data processing was performed in the statistical environment RStudio® applying the TreeLS package that uses the algorithm for automatic detection and the algorithm for determination of the DBH. The number of trees detected at settlement level was 88.4%. It was possible to verify that trees dead and next to the equipment caused considerable shading effects, and that most of the undetected trees were located near the edge of the sample unit. The highest percentage of detection was for the 4x2 spacing (94.3%) possibly due to the lower density of trees. The 3x3 spacing showed the lowest detection rate that can be explained by the presence of cloud noise due to expressive sub-forest. Regarding the diameters, there was a tendency of underestimation in all spacings, however, no significant differences were observed at the 5% level even the algorithm did not detect all the trees in the sample units. It was possible to verify that TLS technology can be applied in the automatic detection and determination of diameters, but the sub-forest generates noises in the cloud of points, which must be removed before the processing of the inventory variables.