DRONE APPLIED IN DIGITAL AGRICULTURE

Abstract

INCREASED INPUT COSTS AND THE DEMAND FOR SUSTAINABILITY IN RURAL AREAS REQUIRE IMPROVEMENTS IN THE EFFICIENCY OF RESOURCE USE IN AGRICULTURE. PRECISION AGRICULTURE (AP) CAN BE USED TO MAKE PRECISE TEMPORAL AND SPATIAL DECISIONS, INCREASING APPLICATION EFFICIENCY AND SUSTAINABLE MANAGEMENT. THE OBJECTIVE OF THIS RESEARCH WAS TO EVALUATE SOME APPLICATIONS WITH UNMANNED AERIAL VEHICLE (UAV) IN THE IDENTIFICATION OF PLANTING FAULTS, PLANTING LINES AND COUNTING OF CASSAVA PLANTS. IT ALSO AIMED TO IDENTIFY WEEDS IN A PRE-PLANTING AREA. THE FLIGHTS WERE CARRIED OUT BETWEEN 10:00 AM AND 2:00 P.M., TO AVOID SHADOW EFFECTS GENERATED BY THE SUN'S TILT OVER OBJECTS. A HIGH OVERLAP WAS ADOPTED FOR THE PHOTOGRAPHS USED. THE FLIGHT HEIGHT SET FOR ALL CAMERA MODELS WAS 40 METERS, SINCE THE AREAS TO BE FLOWN WERE SMALL. REGARDING THE PLANT COUNT AND THE MAPPING OF THE PLANTING LINES AND FAILURES, THEY WERE ELABORATED IN QGIS 2.18 THROUGH THE MOSAIC GENERATED BY DRONEDEPLOY. THEY WERE OBTAINED THROUGH THE SOFTWARE VECTORIZATION TOOLS. THE SPACES WERE COUNTED AS FAILURES, WHEN SUCH SPACE WAS ENOUGH TO ALLOCATE A PLANT OR MORE. FOR THE WEED MAPPING, A VEGETATION INDEX (IV) ELABORATED IN QGIS 2.18 WAS USED, IN WHICH THE CLASS OF EXPOSED SOIL AND WEEDS WERE GENERATED AS A FUNCTION OF THE BIOMASS. THROUGH THE APPLICATION OF HERBICIDE WITH THE MAP OF WEEDS, IT IS POSSIBLE TO CARRY OUT THE APPLICATION AT VARIABLE RATE IN ORDER TO MINIMIZE COSTS, SINCE WHERE ONLY SOIL EXPOSED THE APPLICATION IS NOT EFFECTIVE. IN ADDITION, THE MAP ALSO ALLOWS TO INFER THE SITES THAT NEED A LARGER FLOW RATE DEPENDING ON THE AMOUNT OF WEED BIOMASS. REGARDING THE IDENTIFICATION OF PLANTING FAILURES AND PLANT COUNTS, IT ALLOWS TO INCREASE PRODUCTIVITY AND TO CARRY OUT REPLANTING PLANNING.