20.500.12556/DKUM-54309 ZAZNAVA DALJNOVODOV V PODATKIH LIDAR DETECTION OF POWER TRANSMISSION LINES IN LIDAR DATA V tem magistrskem delu predstavljamo novo metodo za razpoznavo daljnovodov v podatkih LiDAR. Ker so ti predstavljeni kot množice linearno razporejenih točk nad terenom, izvedemo zaznavo področij na katerih se ti nahajajo s pomočjo cilindrične transformacije na osnovi morfološkega odpiranja. S pragovnim filtrom zaznamo predele z visokim odzivom, nato pa preverimo linearnost njihove porazdelitve s postopkom lokalnega prilagajanja premic. Slednjega ponovimo dvakrat, pri čemer najprej zaznamo položaje daljnovodov, nato pa razpoznamo točke posameznih vodnikov. V zadnjem koraku metode razpoznamo še stebre daljnovodov. Predlagana metoda je hitra, saj je zmožna obdelati 5 milijonov točk v manj kot 40 sek. Glede na rezultate opravljenih eksperimentov, pa metoda dosega točnost in specifičnost razpoznave nad 60% ter natančnost nad 99%. In the master thesis we present a new method for detecting power transmission lines within the LiDAR data. As they are presented by linearly distributed sets of points above the ground, we detect areas on which they are located using the top-hat transformation based on the morphological opening. Thresholding is then applied in order to extract areas with high responses. Here, locally fitted surfaces are used twice. First they are used for the detect of power transmission lines and then for identifying points representing a single conductor. In the last step of the method, detection of the transmission towers is achieved. The proposed method is reliable and also fast as it can process 5 million points in less than 40 sec. Experiments show that the method reaches the precision and the specificity above 60%, while the accuracy of the method is above 99%. LiDAR razpoznava vzorcev matematična morfologija geografski informacijski sistemi LiDAR pattern recognition mathematical morphology geographic information systems true true false R. Cvirn Slovenski jezik Angleški jezik Magistrsko delo/naloga 2015-08-28 01:42:32 2015-10-14 08:25:01 2022-06-05 03:05:53 2021-05-11 03:25:10 2015 0 [Maribor 0 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 004.93:621.315.1(043.2) 19304214 URN:SI:UM:DK:ZP3WRP0J MAG_Cvirn_Robi_2015.pdf MAG_Cvirn_Robi_2015.pdf 1 DFB867CBD4630B26D0B8E182885C3CA9 051aa2dc754f471894c53a6ad923fd3af3df04f413279506981be9aeb458a2eb 04c2e9d1-7c08-11eb-bb7a-00155d0001ca https://dk.um.si/Dokument.php?lang=slv&id=79673 Fakulteta za elektrotehniko, računalništvo in informatiko 0 0 0