3D-digiteerimise tehnoloogiate võrdlus

Käesoleva lõputöö eesmärgiks oli digiteerida väikesemõõduline hoone õppemakett kolmel erineval viisil ning leida valitud meetoditest tõhusaim. Digiteerimiseks valitud meetodid olid terrestriline laserskaneerimine, fotogramm-meetria ja käsilaserskaneerimine. Valitud seadmed vastavalt Faro Focus S70, fotokaamera Sony a7R II ja Faro Freestyle 3D. Töö jaotati kolme osasse. Esimeses osas anti ülevaade digiteerimise olemusest ja kasutusvõimalustest. Tutvustati laserskaneerimise ja fotogramm-meetria tehnoloogiaid. Teises osas kirjeldati töö praktilist poolt. Kirjeldati digiteeritavat objekti, anti ülevaade kasutatud seadmetest ning tutvustati andmete kogumise protsessi. Andmetöötluse poolelt kirjeldatakse, kuidas punktipilved kokku ja mõõtkavasse pandi. Kolmandas osas analüüsiti praktilise osa tulemusena saadud kolme erinevat punktipilve. Hinnati seadmete ja tehnoloogiate mõõdistustäpsust koostatud punktipilvede kvaliteedi alusel. Kasutati CloudCompare tarkvara, et mõõdistatud punktide alusel luua tasapinnad ning Autodesk Revit tarkvaraga hinnati geomeetrilist õigsust hoone telgede ja korruste loomise täpsuse põhjal. Tasapinna loomise tulemusel CloudCompare’iga saadi parim tulemus terrestrilise laserskaneerimise mõõdistustulemusest, mille puhul keskmine ruuthälve oli 0,4 mm. Fotogramm-meetria ja käsilaserskanneri punktipilvede puhul oli sama näitaja vastavalt 2,2 mm ja 1,1 mm. Analüüsi tulemustes võrreldi punktipilvedelt saadud mõõte füüsilise maketi pealt võetud mõõtudega ning hinnangu andmiseks leiti keskmised ruuthälbed. Tulemustest selgus, et Revitis maketile telgede joonestamiseks parima tulemuse andis fotogramm-meetria (keskmine ruuthälve 2,500 mm), millele järgnesid käsilaserskanner Faro Freestyle 3D (3,304 mm) ja laserskanner Faro Focus S70 (5,972 mm). Korruste joonestamisel saadi parim tulemus käsilaserskanneriga (keskmine ruuthälve 0,707 mm), järgnesid fotogramm-meetria (1,414 mm) ja terrestriline laserskaneerimine (2,828 mm). Igale seadmele leiti kolme erineva ruuthälbe põhjal liitmääramatus. Parima tulemusega oli Faro Freestyle 3D käsilaserskanner, mille liitmääramatuse näitaja oli 3,548 mm. Fotogramm-meetrial oli vastav näitaja 3,634 mm ja Faro Focus S70 laserskanneril 6,624 mm. Järeldusena võib öelda, et Faro Focus S70 laserskanneriga saadi CloudCompare tarkvaraga parim tulemus tasapinna moodustamiseks ning seal olid ka punktipilve lõiked selgesti arusaadavad, kuid Revitis olid Faro Freestyle 3D ja fotogramm-meetria meetodid parema tulemusega. Kui eesmärgiks on koguda andmeid väiksemate objektide mudeldamiseks, hindab autor parimaks valikuks Faro Freestyle 3D käsilaserskanneri.

The topic of the given thesis is Comparison of 3D Digitization Technologies, in which qualities of three different point clouds created with different technologies are compared. Today, it is possible to create a digital 3D-model from almost any room, object or situation. Depending on the purpose, criteria such as the level of detail, whether it must be scalable or the aim is to create a visually good image, must be established. To create a 3D-model, it is necessary to collect the data from the room, which would form for example, a single point cloud. The quality of the compiled model is determined by comparing it with the object to be digitized. One of the best ways to assess the quality would be to take measurements of the object and the point cloud and compare them. In order to measure data from a model, it must be scaled. The aim of the given thesis was to digitize a model of a small building in three different ways and to find the most effective of the chosen methods. The chosen methods for digitization were terrestrial laser scanning, photogrammetry and handheld laser scanning. Selected devices were respectively Faro Focus S70, Sony a7R II camera and a Faro Freestyle 3D. The work was divided into three parts. The first part gave an overview of the nature and possibilities of digitization. Introduced laser scanning and photogrammetry technologies. The second part gave an overview of the practical aspects of the work. The object to be digitized and the equipment used were introduced, and the processes of data collection and data processing were described. In the third part, three different point clouds obtained were analyzed. The measurement accuracy of equipment and technologies was assessed on the basis of the quality of the compiled point clouds. CloudCompare software was used to create planes based on measured points, and Autodesk Revit software was used to evaluate geometric accuracy based on the accuracy of drawing axes and floors. Creating a plane with CloudCompare, the best result was obtained with the terrestrial laser scanning measurement, which had a mean square deviation of 0.4 mm. In the Revit, the best result for drawing axes on the model was with photogrammetry (mean square deviation of 2.5 mm) and the best result was obtained with a handheld laser scanner when drawing the floors (mean square deviation of 0.707 mm). The combined uncertainty was found for each device based on three different standard deviations. The best result came with the Faro Freestyle 3D handheld laser scanner with a combined uncertainty of 3.548 mm. The photogrammetry had a corresponding figure of 3.634 mm and the Faro Focus S70 laser scanner had the figure of 6.624 mm. In conclusion, the Faro Focus S70 laser scanner gave the best result with CloudCompare software for forming a plane and the point cloud sections were clearly understandable, but in Revit, Faro Freestyle 3D and photogrammetry methods were better. If the goal is to collect data to model smaller objects, the author evaluates the Faro Freestyle 3D handheld laser scanner as the best choice.