Naslov
The use of UAV Point Cloud Object-Based Classification in the Agricultural Land Consolidation

Autori
Šiško, Josip ; Tomić, Hrvoje ; Paar, Rinaldo ; Mastelić- Ivić, Siniša ; Vuković, Iva

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Skup
FIG Working Week 2020

Mjesto i datum
Amsterdam, Nizozemska, 10.05.2020. - 14.05.2020

Vrsta sudjelovanja
Demonstracija

Vrsta recenzije
Nije recenziran

Ključne riječi
klasifikacija oblaka točaka, objektna klasifikacija, okrupnjivanje zemljišta
(UAV, point cloud classification, object-based classification, structure-frommotion, land consolidation)

Sažetak
Development of unmanned aerial vehicles and automated processing (structure-from-motion technique) of photogrammetric measurements led to a large number of different spatial datasets, obtained by using various sensors. In order to use them in various land management tasks, they need to be further processed and/or adjusted. Land Consolidation is the agrarian and technical operation that aims to group and collect the segmented and fragmented holdings into one or more rounded holdings to achieve a more rational agricultural production. As the procedure of LC is a comprehensive, long-lasting and expensive operation, it needs careful and responsible planning. These plans should take into consideration the ratio between consolidation costs and benefits from improved conditions for agricultural producers. To ascertain that, it is necessary to determine areas suitable for consolidation and express their qualitative features in a quantitative manner. This database needs to include a cadastral dataset, physical planning documentation, land use, dry stone wall structures, digital terrain model, pedological and other spatial or non-spatial datasets. That makes possible to determine various indicators and measures to be used in a multi-criteria analysis to make a decision on size, coverage, comprehensiveness, costs and benefits as well as other characteristics of selected land consolidation procedure. Although the use of multispectral and hyperspectral sensors allows relatively simple and reliable classification of natural and artificial surfaces, the paper discusses application use of a UAV point cloud object-based classification of the most widespread UAV sensor - imaging sensor in the visible spectrum range. The comparison of classified data obtained by surveying at different height was done at the case study of UNESCO’s World Heritage Site Stari Grad Plain in Croatia.

Izvorni jezik
Engleski

Znanstvena područja
Geodezija

POVEZANOST RADA