The main study objective was to develop robust processing and analysis techniques to facilitate the use of small-footprint lidar data for estimating forest biophysical parameters measuring individual trees identifiable on the three-dimensional lidar surface. This study derived the digital terrain model from lidar data using an iterative slope-based algorithm and developed processing methods for directly measuring tree height, crown diameter, and stand density. The lidar system used for this study recorded up to four returns per pulse, with an average footprint of 0.65 m and an average distance between laser shots of 0.7 m. The lidar data set was acquired over deciduous, coniferous, and mixed stands of varying age classes and settings typical of the southeastern United States (37° 25' N, 78° 41' W). Lidar processing techniques for identifying and measuring individual trees included data fusion with multispectral optical data and local filtering with both square and circular windows of variable size. The window size was based on canopy height and forest type. The crown diameter was calculated as the average of two values measured along two perpendicular directions from the location of each tree top, by fitting a four-degree polynomial on both profiles. The ground-truth plot design followed the U.S. National Forest Inventory and Analysis (FIA) field data layout. The lidar-derived tree measurements were used with regression models and cross-validation to estimate plot level field inventory data, including volume, basal area, and biomass. FIA subplots of 0.017 ha each were pooled together in two categories, deciduous trees and pines. For the pine plots, lidar measurements explained 97% of the variance associated with the mean height of dominant trees. For deciduous plots, regression models explained 79% of the mean height variance for dominant trees. Results for estimating crown diameter were similar for both pines and deciduous trees, with R2 values of 0.62-0.63 for the dominant trees. R2 values for estimating biomass were 0.82 for pines (RMSE 29 Mg/ha) and 0.32 for deciduous (RMSE 44 Mg/ha). Overall, plot level tree height and crown diameter calculated from individual tree lidar measurements were particularly important in contributing to model fit and prediction of forest volume and biomass. / Ph. D.

In recent years the science revolving image matching algorithms has gotten an upswing mostly due to its benefits in computer vision. This has led to new opportunities for photogrammetric methods to compete with LiDAR data when it comes to 3D-point clouds and generating surface models. In Sweden a project to create a high resolution national height model started in 2009 and today almost the entirety of Sweden has been scanned with LiDAR sensors. The objective for this project is to achieve a height model with high spatial resolution and high accuracy in height. As for today no update of this model is planned in the project so it’s up to each municipality or company who needs a recent height model to update themselves. This thesis aims to investigate the benefits and shortcomings of using photogrammetric measures for generating and updating surface models. Two image matching software are used, ERDAS photogrammetry and Spacemetric Keystone, to generate a 3D point cloud of a rural area in Botkyrka municipality. The point clouds are interpolated into surface models using different interpolation percentiles and different resolutions. The photogrammetric point clouds are evaluated on how well they fit a reference point cloud, the surfaces are evaluated on how they are affected by the different interpolation percentiles and image resolutions. An analysis to see if the accuracy improves when the point cloud is interpolated into a surface. The result shows that photogrammetric point clouds follows the profile of the ground well but contains a lot of noise in the forest covered areas. A lower image resolution improves the accuracy for the forest feature in the surfaces. The results also show that noise-reduction is essential to generate a surface with decent accuracy. Furthermore, the results identify problem areas in dry deciduous forest where the photogrammetric method fails to capture the forest.

I området Ren i Bollnäs kommun ska exploatering av ett nytt bostadsområde ske utifrån detaljplan Ren 30:1. Vanligtvis genomförs hydrologisk modellering med nationella höjdmodellen (NH) som underlag inför exploatering. Med hjälp av metoder som skyfallskartering kan risken för översvämning och dess eventuella konsekvenser förutses. NH tillhandahålls av Lantmäteriet och finns tillgänglig som punktmoln och 2 m raster och kan användas som underlag för hydrologiska modelleringar. NH lämpar sig bäst för nationell och regional planering som ska göras översiktligt på grund av punkttätheten som är förhållandevis låg. Syftet med den här studien var att undersöka olika underlag för hydrologisk modellering genom att jämföra NH mot digitala höjdmodeller från data insamlat med unmanned aerial system (UAS) på flyghöjderna 80 och 120 m. UAS-flygningarna utfördes med en drönare av modellen DJI Phantom 4 Pro V2.0 som samlade in höjddata fotogrammetriskt. Bearbetning, georeferering och skapande av höjdmodell utfördes i programvarorna Agisoft Photoscan och ESRI ArcGIS. GNSS med nätverks-RTK användes för inmätning av flygsignaler och kontrollprofiler. Hydrologisk modellering genomfördes med beräkningar av flödesriktning och flödesackumulation på underlag från NH och flygdata insamlat med UAS i ArcGIS. På alla underlag genomfördes skyfallskarteringar i Caesar Lisflood-FP. Eftersom Caesar Lisflood-FP hade en begränsning för hur stora mängder data som kunde hanteras ändrades cellstorleken på rastren till 1 m istället för 0,5 m, NH utvärderades i originalstorleken 2 m.Med UAS erhölls en högre punkttäthet och utifrån det kunde ett mer högupplöst raster på 0,5 m skapas jämfört med NH med en rasterstorlek på 2 m. Kontroll av höjddata från 120 m och 80 m visade att den lägre flyghöjden genererade en höjdmodell med högre noggrannhet. Utvärdering av överlapp vid skyfallskartering och vattendjup visade att data insamlat med UAS från 120 m och 80 m överensstämde med 96,7 % för vattenutbredning och största vattendjupen var 0,51 m och 0,48 m. Mellan NH och UAS-data fanns större skillnader i både vattendjup och utbredning. Slutsatsen i den här studien blev att olika underlag påverkar den hydrologisk modelleringen genom att t.ex. flödesriktning, vattendjup och utbredning avviker beroende på vilket underlag som används (UAS eller NH) men även upplösning och flyghöjd. / In the area Ren in Bollnäs municipality, development of a new residential area shall take place based on detailed development plan Ren 30: 1. Hydrological modeling with the national height model (NH) is usually carried out as a basis for development. Using methods such as mapping rainfall, the risk of flooding and its possible consequences can be predicted. NH is available at Lantmäteriet as point cloud and 2 m grid and can be used as a basis for hydrological modeling. NH is best suited for national and regional planning to be done briefly, because of the point density that is relatively low. The purpose of this study was to investigate different data sources for hydrological modeling by comparing NH against digital height models collected with unmanned aerial system (UAS) at flight heights of 80 and 120 m. The UAS flights were performed using the model DJI Phantom 4 Pro V2.0 which collected height data photogrammetric. Processing, geo-referencing and creation of height model were performed in the Agisoft Photoscan and ESRI ArcGIS software. GNSS with network RTK was used for surveying flight signals and control profiles. Hydrological modeling was carried out with calculations of flow direction and flow accumulation based on NH and UAS data in ArcGIS. Rainfall modeling were carried out in Caesar Lisflood-FP. Since Caesar Lisflood-FP had a limit on how large amounts of data could be handled, the cell size of the raster was changed to 1 m instead of 0.5 m, NH was evaluated in the original size 2 m. A higher point density was obtained with UAS and due to that, a higher resolution grid of 0.5 m could be created compared to NH with a grid size of 2 m. Control of height data from 120 m and 80 m showed that the lower flight height generated a height model with higher accuracy. The evaluation of rainfall mapping and water depth showed that data collected with UAS from 120 m and 80 m corresponded to 96.7% in water extent, and the largest water depths were 0.51 and 0.48 m. Between NH and the two UAS data, there were greater differences in both water depth and extent. The conclusion in this study was that different basis affect the hydrological modeling by for example flow direction, water depth and water spread differ depending on the data source (UAS or NH) as well on the resolution used, and the flight height.

Hydrologiska analyser blir ett allt vanligare tillvägagångssätt för att kunna planera inför eventuella översvämningar. Då naturkatastrofer som dessa kommer öka i framtiden i takt med att temperaturen höjs och nederbörden ökar så har Fagerstakommun påbörjat framtagning av skyfallskartering. Syftet med studien är att undersöka hur resultatet av skyfallskartering och hydrologisk modellering påverkas av dataunderlagetsupplösning. Ytterligare ett mål med arbetet är att undersöka om den nationella höjdmodellen (NH) lämpar sig att användas som underlag vid framtida skyfallskarteringar.Området som berörs av studien ligger beläget i Fagerstas centrala stadskärna och delades uppi ett område nordväst om Bergslagssjukhuset och ett i stadsdelen Västanfors. Fagersta kommun har genomfört en laserskanning som ligger till grund för dataunderlagen i analyserna. Laserpunkterna harförarbetats i Cyclone 3DRochberäknatsi FME Workbenchgenom TIN (Triangulated Irregular Network) och sedan har digitala höjdmodeller (eng. digital elevation model, DEM) skapats i upplösningarna 0,5 m, 1 m, 2 m, 5 m, 10 m, 25 m och 50 m. Även den nyligen framställda NH användes som ett underlag. Denna består av en DEM i upplösning 1 m. Skyfallskartering genomfördes i ArcMap 10.8 medmetoden att lokalisera lågpunkter som riskerar att svämmas över. Hydrologisk modellering tillämpades i syfte att framställa flödesvägar som visar var vattnet rinner vid översvämning. Resultaten jämfördes med hjälp av feature agreement statisticssom visade hur bra dataunderlagen överensstämde mot referensupplösningen på 0,5 m. Ytterligare valideringar gjordes mot en skyfallskartering som tillhandahölls av Fagersta kommun. Denna genomfördes år 2017 av SMHI på uppdrag av Länsstyrelsen och jämfördes mot 0,5 m, 1 m och NH 1 m.Resultat från skyfallskarteringen visade att NH överensstämmer väl med Fagerstas egna skyfallskartering. Det visade att utbredningen generellt ökar och fill-up värdet minskar vid en lägre upplösning. Tvärsnitt från lågpunkterna visade attdetta även generaliserar djupet. De hydrologiska modelleringarna styrker att lägre upplösning generaliserar området. En slutsats kan dras att upplösningarna på 10 m och uppåt inte böranvändas vid en mer noggrann analys. Däremot kan de vara tillräckliga som underlag vid översiktlig kartering. Ytterligare slutsats drogsatt NH lämpar sig som underlag vid skyfallskartering för de flesta ändamålen. Ett underlag med tätare punkttäthet bör dockanvändas vid analyser som kräver ett mer detaljerat resultat. / Hydrological analyses are becoming a more common tool for being able to prepare for eventual floods. As natural disasters such as these will increase in the future as higher temperatures and increased precipitation Fagersta municipality has begun producing cloudburst maps. The aim of this study is to investigate the effects that resolution has on hydrological modelling and cloudburst mapping. Another goal of the study is to examine if the Swedish national height model (NH) is suitable for use in forthcoming cloudburst analyses.The study area is located in the city center of Fagersta and is divided into two areas, one northwest of the hospital named Bergslagssjukhuset and the other over the district Västanfors. Fagersta municipality has carried out a laserscanning which is used to create the different data sets in the analyses. The measured points have been preworked in Cyclone 3DR and calculated in FME Workbench as a TIN (Triangulated Irregular Network). Digital elevation models (DEM) were later on created with the following resolutions;0,5 m, 1 m, 2 m, 5 m, 10 m, 25 m and 50 m. The recently produced NH was also used as a data set. It consists of a DEM with a resolution of 1 m. The cloudburst mapping was implemented in ArcMap 10.8 with the method to locate blue spots that run the risk of being flooded. Hydrological modelling was applied with the aim of producing stream networks which presents where water will flow in case of a flood. The results were compared by using feature agreement statistics which showed how comparable the data sets were to the reference resolution of 0,5 m. Additional validations were made by using Fagersta municipality's own cloudburst map which was carried out in 2017 by SMHI on behalf of Länsstyrelsen. The datawas compared to 0,5 m, 1 m and NH 1 m.Results from the cloudburst mapping showed that NH compares well with Fagerstas data set. It also showed that the spread of blue spots generally increases and the fill-up valuedecreases at a lower resolution. Cross sections from the blue spots showed that this also generalizes depth. The hydrological modellings strengthens that lower resolutions generalizes areas. A conclusion can be drawn that the resolutions of 10 m and up should not be used in a precise analysis. They can however be used for easy-reference mapping. Another conclusion was drawn that NH is suitable for use as a data set in cloudburst mapping for most purposes. A data set with a better point density should however be used in analyses which require a more detailed result.

Recently, products of Unmanned Aerial System (UAS) integrated through SIFT algorithm and dense cloud matching using structure from motion has gained prominence with tree-level inventory maintenance in forestry. Various studies have been carried out by using UAS imagery to quantify and map forest structure of simple coniferous stands. However, most of the previous works employ methodologies that require manual inputs and lack of reproducibility to other forest systmes. Manual detection of trees and calculation of their attributes can be a time-consuming and complicated process which can be overcome with an automated technique applied by forest managers and/or landowners is highly desired to take full advantage of the readily available UAS remote sensing images. This study presents a methodology for automated measurements of tree height, crown area and crown diameter of hardwood species using UAS images. Different UAS platforms were employed to gather digital data of two hardwood plantations at Martell, Indiana. The resulting aerial images were used to generate the Digital Surface Model (DSM) and Digital Elevation Model (DEM) for the forest stand from which the Crown Height Model (CHM) was derived. The canopy height model can be inputted to the web platform deployed through shiny server (https://feilab.shinyapps.io/Crown/) to derive individual tree parameters automatically. The results show that this automated method provides a high accuracy in individual tree identification (F-score> 90%) and tree-level measurements (RMSEht<1.2m and RMSEcrn<1m). Moreover, tree-level parameter estimation for 4,600 trees were calculated in less than 30 minutes based on a post-processed DSM from UAS-SfM derived images with minimal manual inputs. This study demonstrates the feasibility of automated inventory and measure of tree-level attributes in hardwood plantations with UAS images.

Vilken transportväg är den miljövänligaste för att ta sig från plats A till plats B? För att ta reda på detta krävs det någon form av modell för vägnätet som gör det möjligt att beräkna den totala bränsleförbrukningen för en viss resväg. Ett sätt att skapa en sådan kostnadsmodell är genom att applicera vägnätet på en höjdmodell med mål att vägnätet ska få höjddata. Problemet med denna metod är dock att stora plötsliga höjdskillnader genereras vilket inte motsvarar verkligheten. Syfte med detta arbete är därför att undersöka orsaken eller orsaker till varför stora höjdskillnader uppstår vid en direkt applikation av vägnät på en höjdmodell. I denna studie motsvaras vägnätet av NVDB (Sveriges nationella vägdatabas) som innehåller Sveriges vägar, och höjdmodellen av en höjdmodell framtagen med hjälp av Lantmäteriets laserskanning. Rapporten delas in i två delar - en teoridel och en databearbetning. Den första delen, teoridelen, ska ge en inblick i hur NVDB och Lantmäteriets höjdmodell är framtagna. I denna del finns även information om de tre studieområdena Göteborg, Kiruna och Värnamo. Den andra delen som är databearbetning handlar istället om att slå ihop de två dimensionerna för att senare kunna identifiera eventuella problemområden. För att åstadkomma detta används bland annat Extract By Mask och Focal Statistics i programmet ArcGIS. Resultatet från denna studie visar att den vanligaste orsaken till att problemområden uppstår är att vägnätet i NVDB hamnar utanför den fysiska och verkliga vägen på höjdmodellen. Andra orsaker som kunde identifieras är kritiska punkter såsom tunnlar och broar. Att vägnätet hamnar utanför beror på att generaliseringsregler tillämpas vid framtagning av NVDB. För att lösa problemet med plötsliga höjdförändringar skulle exempelvis interpolation kunna användas för att jämna ut skillnaden. / Which transport route from point A to point B is the most environmentally friendly? To determine it, a model of the road network is needed to calculate the total fuel consumption for a specific route. One way to create this type of cost model is to apply the road network to a height model, intending to assign height data for the road network. However, the problem with this method is that large, sudden height differences occur which do not correspond to reality. The motive of this work is to investigate the cause or causes of why these height differences appear with the direct application of a road network onto a height model. In this report, the road network is NVDB (National Road Database of Sweden) which contains all roads of Sweden, and the height model is a height model produced by the laser scanning of Lantmäteriet. The report is divided into two parts - literature study and data processing. The first part, the literature study, is supposed to give an insight into how NVDB and the height model are created. This part also contains information about the three study areas Göteborg, Kiruna and Värnamo. In the second part, which is the data processing, the two dimensions are merged and are later analyzed to identify any problem areas. To achieve this, the software program ArcGIS is used, and Extract By Mask and Focal Statistics are some used functions. The results from this study indicate that the most common source of problem areas where height differences occur is that the road network in NVDB is placed incorrectly and therefore does not correspond to the physical world whereas often is the road on the height model. Another common source are critical points such as tunnels and bridges. When producing the NVDB, rules of generalization are applied, resulting in the road network placed outside the physical road. A potential solution for the sudden height change is using interpolation between points to level out the height difference.

Segmentation of an image into individual tree crowns is a key step in the processing of remotely sensed data for forestry practice. The doctoral thesis gives a broad overview of this topic. It comprehends theoretical context from mathematical point of view and defines basic terms from airborne imaging and laser scanning. Mathematical methods of tree detection are focused on a robust adaptation to the actual conditions in a region of interest. A novel approach of crown area delineation is introduced, it combines a seeded region growing technique with an active contour as a crown boundary representation. The parametrisation of all algorithms is analysed in a practical half of the thesis and more application-oriented issues are mentioned. Executable computer programs are attached.