Author(s):
Penedos, Pedro Pais
Date: 2018
Persistent ID: http://hdl.handle.net/10362/33277
Origin: Repositório Institucional da UNL
Subject(s): Geographical Information Systems; Geospatial Data Mining; Geostatistics; Green Normalized Difference Vegetation Index; Image Classification; ISO Cluster Analysis; Kriging Interpolation; Multispectral Sensors; Normalized Difference Red Edge Index; Normalized Difference Vegetation Index; Precision Agriculture; Remote Sensing; RGB Imagery; Spatial Autocorrelation; Spatial Variability; Unmanned Aerial Systems; Variable Rate Technology; Vegetation Vigor
Description
Dissertation submitted in partial fulfilment of the requirements for the degree of Master of Science in Geospatial Technologies
It is becoming more present in agriculture’s practices the use of Unmanned Aerial Systems with sensors capable of capturing light, in the visible and in longer wavelengths of the electromagnetic spectrum once reflected on the field. These sensors have been used to perform Remote Sensing also in other knowledge fields, describing phenomenon without the risk, cost and the time consuming processes associated with in site samples collection and analysis by a technician or satellite imagery acquisition. The Vegetation Indexes developed can explain the vigor of the cultivation and its data collection processes are more cost and time efficient, allowing farmers to monitor plant grow in every critical stage. These Vegetation Indexes started by being calculated from satellite and airborne imagery, one of the main source for crop management tools, however UAS is becoming more present in Precision Agriculture, achieving better spatial and temporal resolution. This gap in spatial resolution when studying low density cultivations like olive groves and vineyards, creates Vegetation Index’s maps polluted with noise caused by the soil and therefore difficult to interpret and analyse. Hence, when the agriculture has spaced and low density vegetation becomes challenging to understand and extract information from these vegetation index’s maps regarding different spatial variability patterns of the tree canopy vigor. In these cases, where vegetation is spaced it is important to filter this noise. A Classification Model was developed with the objective of extracting just the vegetation’s canopy data. The soil was filtered and the canopy data interpolated using spatial analysis tools. The final interpolated maps produced can provide meaningful information regarding the spatial variability and be used to support decision making, identifying critical areas to be intervened and managed, or be used as an input for Variable Rate Technology applications.
