Določevanje vrst drevja v različnih valovnih dolžinah aerolaserskih podatkov : magistrsko delo

Lavrič, Maja (author), Kosmatin Fras, Mojca (mentor), Triglav Čekada, Mihaela (comentor)

V gozdarstvu je tehnologija aerolaserskega skeniranja že uveljavljena za različne namene. Podatki aerolaserskega skeniranja nam poleg položajnih koordinat podajo tudi odbojnost površine na kateri je prišlo do odboja laserskega žarka. Odbojnost površine opiše intenziteta vrnjenega odboja, na podlagi katere so različni avtorji že poskušali klasificirati drevesne vrste. S primerjavo povprečnih vrednosti intenzitet posameznih drevesnih vrst smo ugotavljali značilne vrednosti, ki bi pripadale določeni drevesni vrsti. V analizo so bili vključeni štirje seti podatkov. Prva dva seta podatkov sta bila zajeta v času olistanosti z enako valovno dolžino, 1550 nm. Prvi set podatkov je bil zajet spomladi , drugi pa pozno poleti. Tretji in četrti set podatkov sta bila zajeta z valovno dolžino 1064 nm, pri čemer je tretji set podatkov zajet v času neolistanosti, četrti pa v času olistanosti. V analizo je bilo vključenih sedeminpetdeset iglatih dreves in šestinpetdeset listnatih dreves, skupaj sto trinajst dreves različnih drevesnih vrst. Znotraj skupine iglavcev smo se osredotočili na smreko in macesen. Med listavci pa smo podrobneje analizirali oreh, javor, jesen in lipo. Ugotovili smo, da je ločevanje med iglavci in listavci bolj zanesljivo v času neolistanosti, saj je povprečna intenziteta iglavcev dvakrat večja od povprečne intenzitete listavcev, to velja predvsem za tretji set podatkov, pri katerem so podatki zajeti z valovno dolžino 1064 nm. Razlike v povprečni vrednosti intenzitet med listavci in iglavci so dobro vidne tudi v času olistanosti. Povprečna intenziteta listavcev je višja od povprečne intenzitete iglavcev, pri čemer so razlike večje pri zajemu podatkov z valovno dolžino 1064 nm (četrti set podatkov), in manjše pri zajemu podatkov z valovno dolžino 1550 nm. Splošna ugotovitev te raziskave je tudi ta, da je intenziteta lubja in lesenih delov dreves nižja od intenzitete iglic in listov. Izkazalo se je tudi, da na osnovi zmanjšanja ali povečanja deleža prvih odbojev v določeni valovni dolžini lahko pri listavcih ugotovimo ali so bili že olistani ali ne. Na ta način smo ugotovili, da oreh in jesen v prvem setu podatkov še nista imela povsem razvite krošnje, v primerjaviz javorjem in lipo.

• Language: Slovenian
• Keywords: geodezija, magistrska dela, aerolasersko skeniranje, intenziteta, intenziteta, valovna dolžina, listavci, iglavci
• Work type: Master's thesis/paper
• Typology: 2.09 - Master's Thesis
• Organization: FGG - Faculty of Civil and Geodetic Engineering
• Place of publishing: Ljubljana
• Publisher: [M. Lavrič]
• Year: 2016
• Number of pages: X, 56 str., 2 pril.
• PID: 20.500.12556/RUL-88499
• UDC: 528.7:582.091(497.4)(043)
• COBISS.SI-ID: 7876449
• Publication date in RUL: 11.01.2017

Secondary language

• Language: English
• Title: Determination of tree species in lidar data acquired with different wavelengths

Abstract

In forestry, the airborne laser scanning is already being used for various purposes. Data gathered from airborne laser scanning besides positional coordinates also provides surface reflectance from which the laser echoes derived. Echoes intensity describes surface reflectance, upon which some authors already tried to classify individual tree species. By comparing average intensities of individual trees species we assessed characteristic values, which could belong to specified tree species. Four sets of data where included into analysis. First two datasets where gathered in leaf on with wavelength of 1550 nm. First dataset was gathered in spring; second set was gathered in late summer. Third and fourth dataset were gathered with wavelength of 1064 nm, in which the third set was gathered in leaf off, but the fourth in leaf on. Fifty-seven coniferous and fifty-six deciduous trees, together hundred thirteen diverse tree species were included into analysis. Within conifers, we point our focus onto spruce and larch. Among deciduous the focus was on walnut, maple, ash and linden. We found out that classification between coniferous and deciduous is much more reliable in leaf off. Average intensity of coniferous is twice the average intensity of deciduous, the more so with third dataset in which 1064 nm wavelength has been used. Difference in average intensities between deciduous and coniferous also stand out in leaf on. Average intensity of deciduous is higher than average intensity of coniferous. Differences are bigger in the forth dataset (1064 nm) and smaller in the first and the second datasets (1550 nm). In general, our research also shows that intensity of bark and wood is lower than the intensity of needles and leaves. It also turns out that according to the decrease or increase proportion of first echoes in the specific wavelength it can be distinguished whether deciduous are leaf on or not. In this way, we found out that walnut and ash in the first dataset have not had yet fully developed crown in comparison to maple and linden.

• Keywords: geodesy, master of science thesis, airborne laser scanning, intensity, wavelength, deciduous, coniferous