20.500.12556/RUL-182982 Analiza plavja z uporabo bližnjeslikovne fotogrametrije in laserskega skeniranja doktorska disertacija Analysis of the floating objects using close range photogrammetry and laser scanning doctoral dissertation Plavajoči objekti, predvsem večjih dimenzij, z odlaganjem na hidrotehničnih objektih pogosto zmanjšujejo njihovo pretočno sposobnost in povečujejo možnost zamašitve in preplavitev mostov, prepustov in drugih vodnih objektov. Zaustavljanje plavja v kombinaciji z rinjenimi plavinami na mostnih odprtinah in prepustih ter iz njih nastale nestabilne strukture pogosto povzročajo dvig gladin gorvodno in povečujejo možnost nastopa poplavljanja vodotoka. Ob tem prihaja do povečanja hidrodinamičnih obremenitev in spremenjenih tokovnih razmer obtekanja, kar vodi v intenzivne erozijske procese in spodjedanje vodnih ureditev. Poznavanje količine plavja ter njegovih lastnosti, gibanja in transporta je ključnega pomena za obvladovanje s plavjem povezanih težav in preprečitev škodljivega delovanja voda. Z razvojem sodobnih merskih tehnik in metodologij se pojavljajo novi pristopi za zaznavanje plavja. Izbira primerne kombinacije merskih senzorjev in metodologije je odvisna od obravnavanega primera. V doktorski disertaciji smo razvili merski sistem in metodologijo za zaznavanje in klasifikacijo plavja ter določitev oziroma oceno prostornine posameznega vzorca ali skupka več vzorcev, povezanih v grozde. Metodologija temelji na uporabi brezkontaktnih merskih metod. Pri laboratorijskem delu poskusa smo meritve izvedli z uporabo 2D laserskega skenerja in digitalnih (industrijskih) fotoaparatov. Prostornino posameznega vzorca pa izračunali na podlagi uporabe metod RANSAC in MNK ter kombinacije meritev prostornine vzorcev nad vodno gladino in izbrane specifične gostote materiala. Rezultate obdelav oblakov točk laserskega skeniranja in fotogrametričnih meritev smo medsebojno primerjali, prav tako smo rezultate obeh metod primerjali tudi z referenčnimi vrednostmi pridobljenimi z Arhimedovo metodo. V terenskem delu doktorske disertacije smo simulirali razmere na območjih lovilnih mrež na hidrotehničnih objektih oziroma na mestih zagozdenja plavja pri posameznih mostnih konstrukcijah, kjer prihaja do zaustavljanja in nabiranja plavja. Meritve smo izvedli z brezpilotnim letalnikom, na katerega je bil nameščen fotoaparat in/ali laserski skener. Prostornino celotnega plavja smo ocenili z uporabo metod strojnega učenja, s katerimi smo v oblaku točk zaznali objekte, jim določili material in posledično pripisali specifično gostoto. Na podlagi pripisane specifične gostote materiala in znane prostornine vzorca nad vodno gladino smo izračunali prostornino celotnega plavja. Floating objects, particularly those of larger dimensions, often accumulate on hydraulic structures, reducing their flow capacity and increasing the risk of clogging and overtopping of bridges, culverts, and other installations. The retention of floating debris, combined with transported bedload at bridge openings and culvert inlets, can create unstable accumulations that raise upstream water levels and heighten the likelihood of flooding, overtopping, and local scouring of hydraulic structures. Understanding the quantity, characteristics, movement, and transport of floating debris is therefore essential for effective management and mitigation of debris-related hydraulic impacts. With the advancement of modern measurement technologies and methodologies, new approaches for detecting and analysing floating objects have been proposed. The choice of an appropriate combination of sensors and analytical methods depends on the specific problem under investigation. In this doctoral dissertation, a measurement system and methodology were developed for the detection, classification, and volumetric estimation of floating debris, applicable to both individual samples and clusters composed of multiple pieces. The proposed approach is based on non-contact measurement techniques. In laboratory experiments, measurements were carried out using a 2D laser scanner and digital industrial cameras. The volume of each sample was determined using the RANSAC method, along with a combination of above-water volume measurements and assigned specific material density. The point clouds obtained from LiDAR and photogrammetric measurements were compared and validated against reference values derived from Archimedes’ principle. In the fieldwork part of the doctoral dissertation, we aimed to simulate the conditions in the areas of debris boom installed at hydraulic structures, as well as floating object jams forming at individual bridge structures, where the accumulation and retention of floating objects occur. Field experiments were performed using an unmanned aerial vehicle equipped with either a camera or LiDAR system. The total volume of floating objects was determined with machine learning algorithms, which identified and quantified material types within the point cloud. Based on the assigned specific material density and the known above-water volume, the total floating objects volume was calculated. doktorske disertacije grajeno okolje terestrično lasersko skeniranje bližjeslikovna fotogrametrija RANSAC analiza plavja ocena prostornine plavja strojno učenje built environment terrestrial laser scanning close-range photogrammetry floating objects analysis floating objects volume estimation machine learning true false false [Ž. Pleterski] Slovenski jezik Angleški jezik Doktorsko delo/naloga 2026-05-30 08:45:08 2026-05-30 08:45:22 2026-06-04 09:42:31 0000-00-00 00:00:00 2026 0 Ljubljana 2026 Ljubljana XXXII, 152 str. 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 528.7:626/627(043.3) 179378 280425475 21065.pdf 21065.pdf 1 378F9566B58BF7287C4395364634666D 9fc11fd80bf71f8a94b22feddaa55788a57ba6a693b5724ef073d6b89d3f8016 c9bbcd99-5bf2-11f1-b0ab-0050569b8976 https://repozitorij.uni-lj.si/Dokument.php?lang=slv&id=234699 Fakulteta za gradbeništvo in geodezijo 0 0 0