Spray and economics assessment of a UAV‑based ultra‑low‑volume application in olive and citrus orchards

Vol:.(1234567890) Precision Agriculture (2020) 21:226–243 https://doi.org/10.1007/s11119-019-09665-7 1 3 Spray andeconomics assessment ofaUAV‑based ultra‑low‑volume application inolive andcitrus orchards JorgeMartinez‑Guanter1· PabloAgüera2· JuanAgüera3· ManuelPérez‑Ruiz1,4 Published online: 6 May 2019 © Springer Science+Business Media, LLC, part of Springer Nature 2019 Abstract Automation is a new frontier in specialty agriculture equipment. Specifically, unmanned aerial vehicles (UAV), machine vision and robotics will increasingly appear in sustainable agricultural systems. The use of small UAVs retrofitted with spraying systems allows precision aerial applications on small targets. These precision applications can result in significant cost savings and reductions in risk to operators during treatments. This paper presents a novel and practical design and development of a small application system capable of being mounted on an unmanned aerial vehicle for agrochemical spraying tasks and an analysis of the quality of the application and economic costs in olive and citrus orchards compared with those of a conventional treatment. Once the equipment had been developed, field trials in super-high-density olive and citrus orchards were undertaken to evaluate the spray deposition efficiency. For comparison with a conventional hydro-pneumatic sprayer, the field tests took into account parameters such as the applied volume rate, spray drift, application time and equipment costs and depreciation. The results obtained indicate that there was a 7€/ha difference in the application costs between the aerial vehicle and conventional equipment. It is hoped that the conclusions of this work will serve as the basis for a debate about the existing legislation governing this type of aerial work, which can be beneficial in specific cases and should be carried out in a safe and legal manner. Keywords Agrochemical application· UAV· Sprayer· Economic analysis * Manuel Pérez-Ruiz [email protected] 1 Universidad de Sevilla. Área de Ingeniería Agroforestal Dpto. de Ingeniería Aeroespacial y Mecánica de Fluidos, Seville, Spain 2 Dronsap, aDivision ofAgrosap, Mairena del Aljarafe, 41927Seville, Spain 3 Dept. ofRural Engineering, University ofCórdoba, Rabanales Campus, 14014Córdoba, Spain 4 Ctra. Utrera Km 1, 41013Seville, Spain 227 Precision Agriculture (2020) 21:226–243 1 3 Introduction In recent years, agricultural UAVs have gained popularity, leading to a paradigm shift in agricultural production (Limnaios 2014). These aerial platforms with computer-based systems for autonomous or semi-autonomous navigation have become relatively affordable tools for widespread use. The wide range of activities that they can carry out in agricultural environments has turned them into increasingly popular tools in agricultural fields. The ability of UAVs to fly at low altitudes carrying airborne sensors, allows data acquisition with both ultra-high spatial and spectral resolutions (Pajares 2015). UAVs also provide easy and fast mission design, reusability, cost-effectiveness and open-source frameworks (Urbahs and Jonaite 2013). Agronomically relevant data can be obtained for tasks such as mapping vegetation cover (Torres-Sánchez etal. 2013), deriving vegetation index data (Agüera Vega etal. 2015; Berni etal. 2009), estimating nitrogen (Zaman-Allah etal. 2015), reconstructing 3D orchards (Nevalainen etal. 2017; Díaz-Varela etal. 2015) and the early detection of diseases using chlorophyll fluorescence (Zarco-Tejada etal. 2012; West etal. 2017). However, the use of UAVs in agriculture is not limited to the imaging or scouting of crop diseases, pests, weeds or water deficits. Recently, UAVs have been used to perform variable rate application tasks such as selective ultra-low-volume herbicide application (Zhang etal. 2016; Giles and Billing, 2015; Huang etal. 2009) or remote aerial controlled seeding and reforesting (Wired Magazine 2015). Although there is widespread concern regarding the negative environmental consequences of phytosanitary products in liquid form (Panneton et al. 2005), fumigation remains the most common form of application due to its low cost and good efficiency (Giles etal. 2008). The implements and technologies used for this type of application have remained relatively unchanged in recent years. At the European level, the Directive on the Sustainable Use of Pesticides 2009/128/EC (OJEC 2009) has motivated research into precision spraying to find an alternative method of applying these phytosanitary products. The use of hydraulic and hydro-pneumatic sprayers is undeniably widespread in conventional applications. However, in recent years, variable-dose spraying and selective applications have emerged as satisfactory techniques in terms of product savings, environmental safety and end-use quality. UAV technology has proven its usefulness in crop protection and vector control not only for observation and detection but also for the precise application of agrochemicals (Giles and Billing, 2015). It is in Asian countries, and especially for extensive arable crops such as rice, wheat or barley, where the most important developments in UAV spraying techniques have been generated in recent years. Compared with capturing remote sensing information, applications with UAVs still present a considerable technical challenge. The payload and spraying equipment exhibit different behaviours and the power demand is significantly higher than when using cameras. In this type of aerial application, multi-rotor UAVs have a number of advantages over fixed-wing UAVs or spray UAVs, such as their flexibility and manoeuvrability in complex flight patterns, the nonrequirement of a large landing/take-off area and the possibility of hovering over specific points on the ground (Zhang etal. 2016). In addition to these aspects of flight performance, the reduced exposure of workers to chemicals and the ability to apply chemicals in spatially variable patterns in a plot make UAV applications attractive from the point of view of precision agriculture. Previous studies with this type of unmanned system for agricultural tasks such as that carried out by Huang etal. (2009), focused on the use of specific implements mounted on UAVs, such as a rotary electrostatic atomizer Ru etal. (2011) or the co-ordination of a fleet of small multi-rotor vehicles (Wang etal. 2013). Most of these 228 Precision Agriculture (2020) 21:226–243 1 3 studies reported a common conclusion considered in the present work: the use of UAVs in spraying applications makes sense when used with ultra-low-volume (ULV) doses. These types of applications are considered key to precision plant protection because they result in a higher percentage of the chemical applied being collected on the target surface than when using conventional spraying (Bals 1970). Three key factors must be considered in these treatments: the optimum droplet size, the application strategy and the possible effect on the environment (Mount 1985). ULV treatments represent a significant risk of pollution by plant protection products, as they require the generation of fine, easily transportable airborne droplets that can affect sensitive adjacent areas. Whereas ULV applications are paramount since they provide maximum efficiency in targeted areas, the use of UAVs in this context still presents some technical challenges, such as the relatively large size of the drops from commercial nozzles, the difficulty in calculating the dose of active material to be applied and the small application area to which they are limited (Sheng etal. 2016). Based on the above considerations, there is a growing interest in the development of application-specific equipment using UAVs. An economic comparison of these equipment with the standard techniques of spraying can reinforce their use in certain situations. The objectives of this study were (1) to develop a low-cost modular spraying system mounted on a UAV (hereafter referred to as Aerial Treatment for a High Orchard System or ATHOS), (2) to study the quality of the application in terms of a droplet and drift analysis and (3) to perform a comparison with a conventional sprayer application in cost-sensitive terms. The results are expected to demonstrate the suitability of the UAVs as tools for lowvolume applications, comparing them in economic and quality terms against conventional equipment in a real use case. Materials andmethods UAV platform To implement a spraying system capable of applying the insecticide treatment, the DJI s1000 + commercial UAV model (DJI, Shenzhen, China) was used for its autonomy and load capacity. This commercial model has a 15min autonomy with a take-off weight of approximately 6kg. An embedded GNSS receiver with sub-metre accuracy was used for navigation and route generation based on predefined waypoints. An ad hoc telemetry system consisting of a radio modem and a ground receiving station provided real-time information on the flight parameters (height above ground, horizontal and vertical velocities, position of the UAV), while an analogue video system allowed real-time images to be obtained from a first-person view. The maximum payload was 5kg, but it was not possible to exceed 11kg of total mass on take-off. The autonomy of the equipment with the mounted sprayer system described below was measured on several flights with the maximum load, establishing anaverage flight autonomy of 14min. Development oftheUAV spraying system The system was designed and built using low-cost materials and considering the payload limitations (Fig.1a). Five elements compose the sprayer unit: the support structure, the tank to house the plant protection product, the nozzle arm or bar, the set of pipes and pump that make up the hydraulic system and the electrical spray control system (Fig.1b). The 229 Precision Agriculture (2020) 21:226–243 1 3 tank, located at the bottom of the UAV, has a capacity of 5l. The modular nozzle bar was designed to allow two spraying configurations: one with four working nozzles with a 250mm spacing between them and another with a single central anti-drift nozzle. Both the assembly and UAV coupling frame structures were designed and manufactured using 3D-printed parts. Transparent PVC pipes with an inner diameter of 6mm were selected for the pipes, while a small, independently supplied electric pump with a voltage of 12V was used to apply within the circuit a maximum operating pressure of 250kPa. The spray control system allows the pump to be driven and its speed to be varied remotely from the UAV remote control station and enables autonomous application in specific areas using pre-established co-ordinates via the electronic system and GNSS. For this purpose, a pulse width modulation (PWM) control system has been designed, in which the radio signal sent from the UAV receiver adjusts the flow rate of the spraying system. This allows the flow rate of the nozzles to vary individually between 0.10l/min and 0.22l/min at the maximum pump range. For the complete characterisation of the equipment, laboratory tests were carried out in which the average flow rate of each nozzle in both configurations was determined at different pump regimes, and the liquid quantities were measured with graduated laboratory test tubes. In addition, the battery discharge curves were determined with respect to the weight and time to characterise the autonomy of the equipment with different levels of payload based on the automatic recording of the telemetry received. Field experiments To demonstrate the reliability and efficiency of the developed system, a treatment with a commercial insecticide that acts as bait for the olive fruit fly (Bactrocera oleae) and common fruit fly (Drosophila melanogaster) was conducted in both olive tree and citrus tree fields. The first field test took place in an experimental exploitation of super-high-density olive trees in facilities on the University of Córdoba campus in Córdoba, Spain (Latitude, 37º 56′ 6.3″ N; Longitude, 4º 43′ 1.1″ W). The olive trees in this orchard measured approximately 2.5m high and were spaced at 1.5m with a 4m row separation. Four olive tree rows (50m long) were selected, over which flights were made with different pre-programmed routes (straight line, flights following a rectangular pattern, etc.), different application patterns and different forward speeds; the treatments were tested at horizontal forward speeds of 2, 3 and 4ms−1 and heights of 0.5, 1 and 2m above the tree canopy Fig. 1 a Application equipment mounted on board the UAV. b Detail of the spray system developed 230 Precision Agriculture (2020) 21:226–243 1 3 (Fig.3). Wind conditions were measured with a precision handheld anemometer. During this first test, the wind speed was 1.1ms−1. The second test was carried out on a commercial orange grove near Seville, Spain (37 30′ 43.13º N; 5 58′ 12.78º W), whose trees had a separation of 4m and 5.5m between rows, with an average height of 2.45 m. During this second test, the wind speed was 2.1ms−1. In this second test, a John Deere 5510N (John Deere, Illinois, US) 67kW tractor and a Hardi Mercury (Hardi, Nørre Alslev, Denmark) atomizer with a deposit capacity of 3000l was also used for treatments comparison. With the conventional equipment, Table 1 Preliminary test setups for tractor + atomizer sprayer application in citrus orchards Test setup Tractor speed (ms−1) Implement working pressure (kPa) Setup 1 V1 = 2.2 P1 = 1200 Setup 2 V1 = 2.2 P2 = 800 Setup 3 V2 = 2.7 P2 = 800 Setup 4 V2 = 2.7 P1 = 1200 Fig. 2 a UAV aerial application using single-nozzle configuration. b Conventional spraying application of commercial product in orange trees using tractor + implement equipment Fig. 3 Image processing of water-sensitive papers. RAW image is first cropped and binarized. Then, the software detects the contours of the elements in the image (droplets) and counts the contours. The final output shows the resulting information in table form 231 Precision Agriculture (2020) 21:226–243 1 3 different treatments were carried out with the usual working speeds and pressures for the application of the commercial formulation, so that different spraying configurations could be obtained, as shown in Table1 below: From these four configurations, it was decided to carry out the final tests using the Setup 1 and Setup 3 combinations, which will be hereafter referred to as Treatment 1 (T1) and Treatment 2 (T2) since, according to the tractor operator’s criteria, they were the most commonly used possibilities for this type of application. Following the recommendations of the product supplier, only the three nozzles in the middle of the section were used (Fig.2b). The hydro-pneumatic sprayer used hollow cone Hardi 1299-12 (Hardi, Nørre Alslev, Denmark) commercial ceramic nozzles, specific for fine spraying on orchard crops. In the UAV equipment, Hardi ISO F-110 standard flat fan nozzles were used. In both application trials, a commercial formulation that acts as bait for the olive fruit fly (Bactrocera oleae) was chosen as the applied ingredient. It is recommended that this insecticide (Spinosad®, Dow Agrosciences), whose active component is 0.024%, be diluted with water at a ratio of 1:10, prepared immediately prior to application, and applied at a rate of 0.25 to 0.5lha−1 in bands covering 25% of the total surface, which allows 1ha to be treated with an application from the spraying system. In the olive and citrus trees, in two non-consecutive rows, two heights (upper and middle crown height) were covered with water-sensitive paper to evaluate the ability to deposit the spray on the tree. To assess the potential drift caused by the ULV application using both spraying systems, tests were conducted under the ISO 22866 standard methodology for field spray drift. Water-sensitive papers were placed on the rows adjacent to the treated trees to capture any drift from both spraying systems (Fig.3). A specific software based on Python’s OpenCV computer vision module (Bradski 2000) was employed to automatically analyse the water-sensitive paper. The results obtained were analysed by using the R statistics software (R Core Team 2013). The droplet population, mean diameter and area covered were measured using this software. After each flight, samples of the water-sensitive paper were collected and analysed in the laboratory. A total of 150 paper samples were analysed automatically for each field test. The software developed initially takes as an input each of the RGB images, on which using the CopyBorder function of the OpenCV library, it detects the edges of the element due to its difference from the background and performs the automatic cropping of the study area. Once this function is executed, a Gaussian Blur is applied as the only pre-processing, and then the adaptive thresholding is computed using the Gaussian method of the OpenCV library, in which the threshold value is the balanced sum of the neighbouring values. This type of thresholding is done because the lighting conditions in all parts of the water-sensitive paper images were not made with structured light and are therefore different. Afterwards, a binary image is generated on which, using the OpenCV ConnectedComponents function, an analysis is performed based on the shape descriptors, in which the elements that do not belong to the background (the spray droplets) are labelled internally, which allows them to be detected and quantified. Finally, the corresponding outputs are stored in a CSV file for analysis (Fig.3). One of the aspects evaluated was the ability of the equipment to effectively apply the product to the crop and to generate relatively uniform coverage while reaching the top and inside-middle parts of the trees. The distribution of droplet populations, the area covered by the droplets per cm2 and the average diameter of the droplets were obtained automatically for every trial test. The measurements of this software were internally validated by comparison with free software available through the USDA, DepositScan (Zhu etal. 2011), with the only purpose of obtaining the degree of confidence of the 232 Precision Agriculture (2020) 21:226–243 1 3 program developed. Although this comparison does not fall within the scope of this paper, it should be noted that similar results to those obtained with DepositScan were obtained in the tests performed. Economic assessment The aerial platform and conventional application equipment were compared, performing the same type of task, taking into account parameters such as the amount of product applied, the hourly output of the spraying equipment and the cost and payback period of the equipment. For the economic evaluation of the use of conventional equipment for the localised application of fly bait, the calculations were based on the technical considerations proposed by the ASABE (American Society of Agricultural and Biological Engineers) and MAGRAMA (Spanish Ministry of Agriculture), collected by Boto etal. (2004). In these guidelines, the costs of agricultural operations were analysed by including the following: (i) fixed depreciation costs, interest, machinery accommodation, insurance and taxes; (ii) variable cost (VCs), including fuel, lubricants, tyre wear, overall repair and maintenance cost (RMCs), and tractor labour; and (iii) the total hourly cost of operation plus tractor and implementation costs. For the economic evaluation of the application, a theoretical working scenario was established on an intensive olive grove plot similar to that described in the Field Experiments Section, with a planting frame of 1.5 × 3m and a number of trees per hectare of approximately 2220. In this scenario, a ULV treatment was proposed, in which both pieces of equipment (the hydropneumatic sprayer and the UAV) must cover 50% of the field area (see Fig.4). To calculate the application costs, a comparative economic study was carried out on the basis of a tractor with the characteristics listed in Table2 and the ATHOS system described above. Based on the preliminary results of the field trials and knowing the design and development costs, an economic assessment of the ATHOS system was carried out with empirical data, and a comparison was made with the previous analysis since no relevant studies or references have been found to date. Fig. 4 a Aerial image of the field test area in olive trees, b diagram of the flight pattern over the olive trees, and c measurement heights where water-sensitive paper was located on all the trees 233 Precision Agriculture (2020) 21:226–243 1 3 Results anddiscussion UAV spraying system Achieving a stable and uniform flight behaviour was one of the main challenges once the spraying components were integrated and mounted on the UAV. A number of adjustments had to be made to the centre of gravity and weight compensation, considering factors such as the possible “swell” of the liquid inside the tank during flight and the progressive decrease in the weight of the assembly during operation. The swell limited the maximum forward speed and manoeuvrability of the assembly, while the weight decrease affected the discharge curve of the battery and the power generated by the engines. To avoid these complications, the variations in the system’s gravity centre during flight were calculated, and the tank support was built so that it was as close as possible to the structure to prevent swaying in the air, which could cause alterations in flight. Regarding the application system, the flow rate of each of the nozzles was tested in the laboratory by collecting samples in graduated laboratory tubes, obtaining an average flow rate of 0.22l/min for each nozzle at maximum pump speed. The results of the battery discharge tests and the evolution of the weight over time are shown in Fig.5. As observed in the curves in Fig.5, the moment of maximum mass (corresponding to take-off and the first few seconds of flight) corresponded to the maximum power demand of the engines and, therefore, to the most marked decrease in the battery charge voltage. In addition, both curves show that once the entire phytosanitary product was consumed (approximately 5min or 300s), the total weight of the system stabilised at approximately 6kg, while the voltage of the battery remained at 21.5V. This battery charge still allowed up to 10 more minutes of flight time (software warnings were created to signal when the Table 2 Characteristics of the tractor used in the economic assessment Rated power 78kW Max power PTR 70kW Acquisition value 30000 € Annual working hours 1000h/year Lifetime period 12years Period 12000h Fig. 5 a Curves of the evolution of the ATHOS equipment’s weight and b the battery discharge over time 234 Precision Agriculture (2020) 21:226–243 1 3 battery voltage reached this point) so that if the equipment is far from the operator, there would still be time for a safe return and landing. Spraying assessment withUAV andconventional equipment In field tests carried out to verify the performance of the aerial application system, different combinations of forward speeds and flight heights were evaluated. Final test conditions of a 3ms−1 horizontal speed and 1m height above the canopy were established. Olive tree spraying The analysis of the water-sensitive papers used in the first tests carried out in the olive grove using ATHOS gave the results shown in Table3 below. For the first results in the methodological approach, the drift results were ignored in the calculations. In addition to the numerical results, it was observed in the field that the air stream generated by the blades of the UAV allowed the penetration of the liquid into the target surfaces inside the tree. Spraying in high-density olive groves is influenced by the structure of the tree, as they are narrow and high in shape. With an aerial spraying system such as the ATHOS, the spray can be applied effectively to the middle part of the tree, one of the objectives of this first field test in the bait product application. Figure6 shows how these droplet populations and their mean diameter were distributed over the two different measurement areas. They include the results obtained with respect to the drift. With regard to the population of drops obtained from this application, it was observed that in the upper part of the olive tree, an average deposition of approximately 25drops/cm2 was obtained, while in the middle part of the tree, an average of 13drops/ cm2 was obtained. In addition, a fairly uniform average drop diameter was obtained in both areas of the tree (144µm in the upper part and 120µm in the middle part). Although these distributions may be scattered, these results are close to what is demanded by the manufacturer of the commercial fly bait product, as only a tree wetting is sought to attract the insect. It is shown that even under the test conditions with a 2.1ms−1 wind, the drift was very low, almost negligible. This fact is a point that will later be presented in the discussion of the results as a benefit of the ULV application equipment. Citrus tree spraying For the second test, where conventional application equipment was available, a comparison was made with the ATHOS spraying equipment. As stated previously in the Field Table 3 Results of the image analysis of 150 water-sensitive paper samples after the ATHOS spraying application on olive trees N = 150 Population (drops/cm2) Coverage (%) Avg. drop diameter (µm) Min 0.1012 0 105 Median 18.2186 2.011 132 Mean 22.9013 4.549 133 Max 76.973 15 206 Std 23.145 3.630 23.234 241 Precision Agriculture (2020) 21:226–243 1 3 With the tractor, the real capacity could be estimated by multiplying the theoretical efficiency by a factor of 0.8, obtaining a value of 4.8hah−1, which implies that application to 1ha can be completed in 12.5min. With the ATHOS system, the real or effective performance factor was increased to 0.9, as turns and manoeuvrability in the field are much faster, requiring less time than those with the tractor. In this way, a real capacity of 5.4hah−1, or 11.1minha−1, could be obtained. Notably, the speed of 2.7ms−1 set for the study could be increased almost twofold when using ATHOS, thus reducing the estimated time needed to treat one hectare. Conclusions The design and construction of a low-cost hydraulic sprayer model (ATHOS) capable of performing specific ultra-low volume applications was successfully completed. Its performance in flight and ability to apply chemicals was field-tested on olive and citrus trees. The droplet population, distribution on canopy, covered leaf area and droplet diameter were measured. In addition, a comparative analysis of the application costs between conventional equipment and the developed ATHOS system was carried out. The following conclusions can be drawn from the present work: Table 5 Summary of tractor + sprayer costs FCs of tractor + sprayer Depreciation cost of tractor 1874.97 €/year Depreciation cost of sprayer 632.66 €/year Cost of interest 562.51 €/year Shedding costs 225 €/year Insurance, taxes 225 €/year Total FCs per year 3520.14 €/year Total FCs per hour 3.52 €/h VCs Fuel consumption 10.29 €/h Lubricant consumption 0.21 €/h Hydraulic oil consumption 0.30 €/h Tyre wear 0.6 €/h Repair/maintenance 1.08 €/h Total VCs per hour 12.48 €/h Labour force Tractor driver (salary + national insurance) 10 €/h Total cost of conventional tractor application 26 €/h Table 6 FCs of the developed ATHOS spraying system Legal declaration operator company/autonomous 300 € Theoretical/practical/medical pilot licence 900 € UAV insurance 200 €/year UAV equipment + remote control 4000 € Developed spraying system 2500 € Four batteries 700 € Total FCs per hour 12.3 €/h 242 Precision Agriculture (2020) 21:226–243 1 3 • The results obtained from the aerial application with the ATHOS system demonstrated that this type of application equipment can perform variable spraying tasks at specific locations in a uniform manner and with an acceptable quality in terms of coverage, population and droplet diameter. In the comparison with the conventional application, the conventional treatment equipment generates an over-population of drops and a wetting that can lead to over-treatment and, therefore, product waste. However, the treatment using the UAV showed uniformity in both the droplet population and diameter, both being sufficiently homogeneous to consider that an effective treatment had been carried out. As shown in the results, a combination of both systems could lead to an optimised treatment in the side and top parts of the tree. • Drift in crop protection applications represents a problem when dealing with conventional equipment, especially in windy conditions. In this work, it is clear that the conventional application equipment, in a normal working setup and in low wind conditions, generated a drift large enough to be an environmental problem. In this respect, the use of UAVs has significantly reduced the drift. This reduction in drift is mainly due to two factors: the lower power of the airflow generated by the flight equipment with respect to the atomizer fan and the direction of the atomizer fan with respect to the structure of the tree. Another important consideration in terms of sustainability is that UAV applications avoid the soil compaction problems caused by machinery. • There was not much difference in the application cost of the aerial platform and conventional equipment (33.80 € ha−1compared to 26 € ha−1, respectively). The costs of the developed equipment were analysed on the basis of estimates from the team’s own experience, to the best of the author’s knowledge, no economic studies of this type of application have been carried out. • The capacity in hectares per hour of equipment such as ATHOS was calculated to be greater than that of conventional equipment because the treatments can be carried out at a higher rate of advance and with greater manoeuvrability. Acknowledgement The authors would like to thank DRONSAP, the UAV division of AGROSAP, for their participation in the design of the equipment and in the trials. We would also like to thank the “World Olive Germplasm Bank” of the University of Cordoba for allowing us to use its facilities. References Agüera Vega, F., Carvajal Ramírez, F., Pérez Saiz, M., & Orgaz Rosúa, F. (2015). Multi-temporal imaging using an unmanned aerial vehicle for monitoring a sunflower crop. Biosystems Engineering, 132, 19–27. Bals, E. J. (1970). The principles of and new developments in ultra-low volume spraying. In Proceedings of the fifth British insecticide and fungicide conference (Vol. 1, pp. 189–193). Alton, UK: British Crop Protection Council. Berni, J. A. J., Zarco-Tejada, P. J., Suárez, L., González-Dugo, V., & Fereres, E. (2009). Remote sensing of vegetation from UAV platforms using lightweight multispectral and thermal imaging sensors. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 38(6), 6. Boto, J. A., Pastrana, P., & Suárez de Cepeda, M. (2004). Consumos energéticos en las operaciones agrícolas en España (Energy Consumption in Agricultural Operations in Spain). Madrid, Spain: Spanish Ministry of Agriculture, Fisheries and Food. Bradski, G. (2000). The opencv library. Dr Dobb’s Journal of Software Tools, 25, 120–125. 243 Precision Agriculture (2020) 21:226–243 1 3 Díaz-Varela, R. A., de la Rosa, R., León, L., & Zarco-Tejada, P. J. (2015). High-resolution airborne UAV imagery to assess olive tree crown parameters using 3D Photo reconstruction: Application in breeding trials. Remote Sensing, 7, 4213–4232. Giles, D., & Billing, R. (2015). Deployment and performance of a UAV for crop spraying. Chemical Engineering Transactions, 44, 307–312. Giles, D. K., Akesson, N. B., & Yates, W. E. (2008). Pesticide application technology: Research and development and the growth of the industry. Transactions of ASABE, 51, 397–403. Huang, Y., Hoffmann, W. C., Lan, Y., Wu, W., & Fritz, B. K. (2009). Development of a spray system for an unmanned aerial vehicle platform. Applied Engineering in Agriculture, 25, 803–809. Limnaios, G. (2014). Current usage of unmanned aircraft systems (UAS) and future challenges: A mission oriented simulator for UAS as a tool for design and performance evaluation. Journals of Computation and Modelling, 4(1), 167–188. Mount, G.A. (1985). Ultra Low Volume Application of Insecticides for Vector Control. Geneva, Switzerland: World Health Organisation, WHO/VBC/85.919 Nevalainen, O., Honkavaara, E., Tuominen, S., Viljanen, N., Hakala, T., Yu, X., etal. (2017). Individual tree detection and classification with UAV-based photogrammetric point clouds and hyperspectral imaging. Remote Sensing, 9(3), 185. OJEC. (2009). Directive 2009/128/EC of the European Parliament and of the Council of 21 October 2009 establishing a framework for community action to achieve the sustainable use of pesticides (Text with EEA relevance) Pajares, G. (2015). Overview and current status of remote sensing applications based on unmanned aerial vehicles (UAVs). Photogrammetric Engineering & Remote Sensing, 81, 281–330. Panneton, B., Lacasse, B., & Piché, M. (2005). Effect of air-jet configuration on spray coverage in vineyards. Biosystems Engineering, 90, 173–184. Sheng, W., Yubin, L., Jiantado, Z., Shenghua, L., Haiyan, Z., & Hang, H. (2016). Analysis and experiment on atomization characteristics of ultra-low-volume swirl nozzle for agricultural unmanned aviation vehicle. Transactions of the Chinese Society of Agricultural Engineering, 32(20), 85–93. R Core Team (2013). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R-proje ct.org/. Torres-Sánchez, J., López-Granados, F., De Castro, A. I., & Peña-Barragán, J. M. (2013). Configuration and specifications of an unmanned aerial vehicle (UAV) for early site-specific weed management. PLoS ONE, 8(3), e58210. Urbahs, A., & Jonaite, I. (2013). Features of the use of unmanned aerial vehicles for agriculture applications. Aviation, 17, 170–175. Wang, Z., Lan, Y., Hoffmann, W.C., Wang, Y., & Zheng, Y. (2013). Low altitude and multiple helicopter formation in precision agriculture. Paper no 131618681, St Joseph, MI, USA: ASABE. West, J. S., Canning, G. G. M., Perryman, S. A., & King, K. (2017). Novel Technologies for the detection of Fusarium head blight disease and airborne inoculum. Tropical Plant Pathology, 42(3), 203–209. Wired Magazine (2015) Re-Planting a Forest, One Drone at a Time. Retrieved August, 2018, from https :// www.wired .com/brand lab/2015/07/re-plant ing-fores t-one-drone -time/ Zaman-Allah, M., Vergara, O., Araus, J. L., Tarekegne, A., Magorokosho, C., Zarco-Tejada, P. J., etal. (2015). Unmanned aerial platform-based multi-spectral imaging for field phenotyping of maize. Plant Methods, 11, 35. Zarco-Tejada, P. J., González-Dugo, V., & Berni, J. A. J. (2012). Fluorescence, temperature and narrowband indices acquired from a UAV platform for water stress detection using a micro-hyperspectral imager and a thermal camera. Remote Sensing of Environment, 117, 322–337. Zhang, P., Deng, L., Lyu, Q., He, S. L., Yi, S. L., Liu, Y. D., etal. (2016). Effects of citrus tree-shape and spraying height of small unmanned aerial vehicle on droplet distribution. International Journal of Agricultural & Biological Engineering, 9(4), 45–52. Ru Y., Zhou, H., Fan, Q., & Wu, X. (2011). Design and investigation of ultra-low volume centrifugal spraying system on aerial plant protection. Paper No. 11-10663, St Joseph, MI, USA: ASABE. Zhu, H., Salyani, M., & Fox, R. D. (2011). A portable scanning system for evaluation of spray deposit distribution. Computers and Electronics in Agriculture, 76(1), 38–43. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.