GEOSPATIAL ANALYSIS USING LIDAR IN FORESTRY AND UAV PHOTOGRAMMETRY IN URBAN REGIONS

Abstract

The recent advancements in Close-Range Remote Sensing techniques are creating opportunities to explore more about the forest ecosystem more precisely. Remote Sensing of forest landscapes is sometimes quite challenging, considering the terrain's ruggedness, remoteness, and presence of dense canopy structure. The Light Detection and Ranging (LiDAR) scanners are a boon to forest applications because they provide detailed and precise information about the forest structure from ground to canopy top level in 3D. However, if research is focused in a close confined region that needs dense information, a Terrestrial Laser Scanner (TLS) proves to be an efficient way of acquiring the information. TLS is an instrument used to acquire data from the ground surface level. The point cloud data collected with TLS is processed to get accurate and precise information at the ground level, under the canopy, and also in tree canopy regions due to its high penetration capability. Under this research theme of LiDAR in forestry, when a TLS survey is done, scan positions and the number of scans are essential parameters to consider before acquiring datasets of forest plots. Depending on the forest plot size, tree density, and structure, multiple TLS scans are acquired to cover the forest plot from all directions to avoid any voids in the dataset that are generated. However, while increasing the number of scans, one often tends to increase the data redundancy as one keep acquiring data for the same region from multiple scan positions. Which makes it necessary to evaluate the efficiency of different TLS combinations. In this research, forest plots of 25m x 25m with different tree densities is considered. A total of nine scans are acquired in each forest plot, and from these nine scans, six different combinations were made to evaluate the 3D vegetation structure derived from each scan combination, such as Center Scans (CS), Four Corners Scans (FCS), Four Corners with Center Scans (FCwCS), Four Sides Center Scans (FSCS), Four Sides Center with Center Scans (FSCwCS), and All Nine Scans (ANS). An analysis is carried out to determine the optimum number of TLS scans and positions needed to sufficiently generate the number of ground points to generate a Digital Terrain Model (DTM). The relative DTM assessment was done between all five combinations and the ANS combination and observed that the FSCwCS combination is more efficient for optimum DTM generation. Similarly, an extensive qualitative analysis was carried out to examine the capability and efficiency of TLS combinations to generate canopy top points in each combination. The quantification of tree canopy top points obtained from a TLS point cloud is crucial as the canopy top points are used to derive the Digital Surface Model (DSM) and Canopy Height Model (CHM). I also performed a statistical evaluation by calculating the differences in the canopy top points between ANS and five other combinations. I found that the most significantly different combination was FSCwCS respective to ANS. The Root Mean Squared Error (RMSE) of the deviations in tree canopy top points obtained for plots TLS_Plot1 and TLS_Plot2 ranged from 0.89 m to 14.98 m and 0.61 m to 7.78 m, respectively. The relative Root Mean Squared Error (rRMSE) obtained for plots TLS_Plot1 and TLS_Plot2 ranged from 0.15% to 2.48% and 0.096% to 1.22%, respectively. Further, it is also crucial to segregate the unstructured 3D forest point cloud data into various classes for accurate information extraction of each feature in the forest, such as stem, canopy, terrain, dead wood, etc. The Forest Structural Complexity Tool (FSCT) algorithm is one such algorithm that has been recently developed for forest point cloud classification. So, I focused on the classification function of this tool. The classification function is classifying the forest point cloud data into stems, leaves, some branches, wood debris, lying dead wood, etc. I tried to investigate the accuracy of stem classification using the FSCT algorithm. However, a qualitative assessment must be done to evaluate the accuracy of the point cloud classification obtained for different tree densities. For this reason, stems from both the forest point clouds were manually extracted and compared with the stem points extracted from the FSCT algorithm classification, and the accuracy was evaluated. The FSCT algorithm-based stem classification achieved accuracy for TLS_Plot1a and TLS_Plot2a is 94.80% and 96.28%.