Browsing by Subject "UAV aerial imagery"

The interest to use drones in agricultural field monitoring is increasing because drones have become more popular during last few years. This is due to the affordable price of them, multi-function features in different conditions, and easy usability. The aim of this study was to explore the use/utilization of drone-multispectral camera system in Finnish field crop production. The aim was also to study if near-infrared map and normalized difference index (NDVI) can be used as a biomass map of grass and is it possible to identify how useful drones are as a supplementary visual aid in field inspection of the crop. Field tests were conducted on the Koirasuo field plot of the Viikki research farm and on the Isokytö field plot in Töysä in Southern Osrtobothnia. The test fields were mapped with drone-multispectral camera system. Reference samples crop were collected and field measurements were made based on these maps. The reference samples and field measurements made were: harvesting time sample, measurement of crop length, leaf area index and chlorophyll content. General crop and plot observations were made also. These results were compared with biomass samples. During the data collection period the Downwelling Light Sensor (DLS) i.e. the camera’s auto-adjusting brightness sensor, caused significant distortions in the maps. More than half of material was useless for in this study for this reason. Useless pictures reflection value was different than normal conditions. However, a solution was found, and photoshooting could be completed. Based on the all results near-infrared maps (R²= 55) and NDVI index (R²= 22) can be used to do only rough biomass estimates on the grass field. Maps can be used also to identify low biomass areas on the field and to define the locations where more measurements or observations are needed. The same reflection from the crop can results from several factors, such as bare ground and dense vegetation. Without reference material, vegetation maps are difficult to analyse. Based on the results, chlorophyll content varies irregularly in different areas mixed grass corp. Maps can be used to explain the history of land use, to detect the locations on piped ditches/sup surfaces drains. Instead it is challenging to identify individual weeds because crop pixel reflection value depends several factors and weeds and crop can be in same pixel area. Pixel accuracy depend flight height and camera resolution. This thesis used 50 meter flight height and spatial resolution was 3,3 cm.