| Resumo: | In continental Portugal, forest fires are considered the biggest and most serious cause of forest deterioration and therefore the introduction of forest management mechanisms and biomass monitoring are imperative for a better future. However, conducting field studies on a large scale is a very expensive and time-consuming task. Alternatively, through remote sensing via a LiDAR, it becomes possible to map, with high accuracy, forest parameters such as tree height, diameter at breast height or tree canopy length in order to carry out other relevant estimates such as above ground biomass. In this sense, this dissertation aims to develop a system capable of, through algorithms and filters of point cloud processing, as statistical outlier removal, progressive morphological filters and region growing segmentation, extract in detail,a digital terrain model and correctly detect the number of trees in a given area, proceeding to the measurement of some interesting variables from the point of view of a forest inventory. Thus, testing data of different characteristics, our detection method obtained positive results, with all the average detection rates above 80 %.
|
|---|