Introduction

The interest in and use of agricultural aviation with a focus on aerial application (especially pesticide application) by unmanned aircraft systems (UAS) have increased in China in recent decades (Zhou et al. 2013). UAS (or unmanned aerial vehicles (UAV) or drones) have been widely used in agriculture, especially precision agriculture (Zhang and Kovacs 2012; Huang et al. 2013; Shahbazi et al. 2014), in various countries. Crop health monitoring through aerial imagery is among the most popular UAS applications in precision agriculture (Zhang and Kovacs 2012; Shahbazi et al. 2014).

Although aerial application was adopted in China in the 1950s (Luo 2014), its use on private farms was limited. Its application was mostly on state-owned farms in agricultural reclamation areas in northeastern and northwestern parts of the country, and was generally aerial spraying by administrative departments for large-scale pest control (Zhang et al. 2014; Zhang 2011). Complicated terrain, small farm size with separated plots, high cost, poor availability and accessibility of aerial vehicles limited the use of aerial application in China. Mountains, high plateaus and hills comprise ~69% of the total area in the country (Yang et al. 2002). The average farm size in China is amongst the smallest in the world (0.67 ha compared with 178.35 ha in the United States), and more than half (57.5%) of the farms are small (<2 ha) (Eastwood et al. 2010). The percentage of small farms is also the largest of any country (Thapa 2009). Most farms are in mountainous and hill areas, away from roads. Even farms in plain areas are usually separated by irrigation ditches and small paths, typically with trees and power poles alongside. These have discouraged the agricultural use of large aerial vehicles because of a lack of operational space for loading, takeoff and landing. Furthermore, aerial applications mostly used manned aircraft, and their cost was usually not affordable by general farm owners.

However, the situation has changed and agricultural use of UAS has been rising in China during the past two decades. Increased interest in UAS was driven by the growing need to improve farming efficiency because of the decline in labour force in rural areas caused by urbanization and rural–urban migration (Liu et al. 2003). Operational farm size has increased with the growth of agricultural co-operatives, which are a form of producer association that provide inputs to their members or market their members’ output (Sonntag et al. 2005; Deng et al. 2010). Land leasing and contract farming also contribute to the increase of operational farm size. Operational farm size is likely to continue increasing with support from a new policy on rural land transfer and appropriately scaled operations for farming. Declared in November 2014, this policy was aimed at facilitating the right to transfer rural land use to agricultural purposes and promote larger-scale agricultural industry. This leads to increasing demand for accuracy and efficiency in farming. Furthermore, action plans for zero growth in fertilizer and pesticide application by 2020, announced by the Ministry of Agriculture in February 2015, project the need for precise and efficient fertilizer and pesticide application. More than 88% of pesticide sprayers are manually operated in China (China Agriculture Yearbook Editorial Committee 2016), and so labour force shortages have led crop growers to start resorting to aerial application for crop management.

Concerns over the health of farm workers and the environment have also promoted interest in UAS applications in China. To reduce harmful runoff to the environment and the threat of pesticides to farmers, an increasing number of agricultural administration bureaus and village chiefs, especially those in the developed southeastern parts of the country, are more willing to accept and adopt new technologies and to use UAS in crop management. The latter users (local governments) collectively purchase UAS and provide aerial application services to all growers in their administrative areas. Owing to the farming conditions described above, most UAS users prefer relatively small and low-cost UAS over manned aerial vehicles, and rotary-wing UAS over fixed-wing UAS. These small UAS are flexible to operate in complicated terrain situations, e.g., farms with trees, electric poles, irrigation ditches and small paths between them (Fig. 1).

Fig. 1
figure 1

Four working UAVs. a YT-A10 by Yinlong Energy Co. Ltd. (weight 8.7 kg, payload 5 kg) spraying mango plantation in Guangxi province; b SW1800ZB by Swiftly UAV Co. Ltd. (weight 11 kg, payload 12 kg) spraying plantain in Guangxi province; c JF-DB5 by Jifengchun Hangkong Co. Ltd., (weight 5.5 kg, payload 5 kg) spraying rice in Zhejiang province; d JF-Y10 by Jifengchun Hangkong Co. Ltd. (weight 20 kg, payload 10 kg) spraying rice in Liaoning province. (photos c and d courtesy by Jinfengchun Hangkong Co. Ltd., Shenyang, Liaoning China)

Full size image

Currently in China, reported interest in agricultural use of UAS focuses on aerial application, crop management and crop protection materials. UAS are also operated as a remote sensing tool to acquire spectral images by onboard sensors, and as “photographers” with consumer cameras to take digital photos. The spectral images and digital photos are used for crop classification, plant status assessment, information extraction from cultivated land and invasive species monitoring. In the following sections, current applications of UAS to agriculture, especially in plant protection, are discussed and summarised, with an outlook for future use of UAS in China.

Aerial application

Aerial applications, especially for crop protection, constitute the major use of UAS in Chinese agriculture. Optimal UAS working parameters for best aerial spray droplet deposition, UAS positioning and navigation, and variable rate spraying technology are fundamental to precise aerial spraying and of great interest.

Efficiency of aerial application

In China, small UAS with payload capacity between 10 and 25 kg are generally used for aerial applications (Liu et al. 2014). The payload capacity within this range is 70.5% in a dataset of 237 UAS models from 104 manufacturers (Fig. 2). Field tests for coverage rate and spraying performance were conducted for UAS aerial application, with various flight speeds, altitudes and spray swath settings. From the literature, aerial pesticide application with UAS in China can achieve a coverage of 4–10 ha/h. This is 30–100 times more than manual spraying and 4–33 times more than a ground-based crop sprayer, with associated substantial energy savings and income increase (Table 1). The efficiency of an unmanned helicopter with 20-kg payload is three to four times that of a tractor-mounted sprayer and 30 times that of a manual sprayer (Liu et al. 2014). Compared with ground machinery, aerial application could reduce the cost by 105 CNY (USD 15Footnote 1) ha−1 (Zhou et al. 2013). In a plant hopper control test, the 20-kg payload UAS HY-B-15L (Gao Ke Xing Nong Co. Ltd., Shenzhen, Guangdong, China) had a 7.5 times faster spraying rate while still having greater insecticidal efficiency and persistence period than a three-person, hand-operated lance stretcher sprayer (Qin et al. 2016).

Fig. 2
figure 2

236 UAS models from 104 manufacturers in China grouped by UAS type, power type and payload capacity (one fixed wing UAS with 100 kg payload capacity was not included into the figure; payload capacities of all three gas powered multi-rotary wing UAS models are greater than or equal to 15 kg)

Full size image
Table 1 Efficiency improvements from the literature for agricultural UAS aerial application in China
Full size table

Compared with UAS efficiency improvement results from the literature for UAS aerial application in China (Table 1), specifications of the 237 UAS model dataset claimed to achieve application coverage rates from 1.33 ha/h (UAS model 3WDM8-05 by Yuren Agricultural Aviation Co., Ltd., Zhuhai, Guangdong, China) to 120 ha/h (UAS model DB-9 by Linhang Unmanned Aerial Vehicle Co., Ltd., Zhengzhou, Henan, China). These rates are 10–1 200 times higher than manual spraying and 1.33–396 times higher than a ground-based crop sprayer.

Currently, more than 60% of agricultural UAS in China have payload capacities less than 15 kg (Fig. 2). Generally, a UAS sprayer with larger payload capacity can have more spraying efficiency than a smaller-capacity UAS if it can be operated at greater flight speed and spraying swath. With increasing farm size, the demand for >15 kg payload capacity and longer flight duration of UAS increases. As hybrid cars have brought higher fuel efficiency, a hybrid UAS could be a solution to the need for larger payload and longer flight duration. Hybrid UAS for plant protection has been of interest. For example, a power matching experiment was conducted for hybrid coaxial rotors and multi-rotor UAS’ power system. The result could satisfy the need for aerial application (Lu et al. 2016).

Optimal UAS operational altitudes and speeds for aerial spraying

Various UAS models have been tested under different flight parameter settings to find optimal working altitudes and speeds that could achieve optimum spray droplet deposition for UAS spraying on many crops (Table 2 is a summary of spray droplet deposition test results from the literature for UAS spraying in China). Spray droplet deposition distributions of various flight speed and altitude combinations were evaluated in rice fields by an infrared thermal imaging system (VarioCAM hr inspect 480, InfraTec GmbH, Dresden, Germany) for a WPH642 unmanned helicopter. Optimal flight altitude and speed were 2 m and 1.5 m/s, respectively (Zhang et al. 2012). A similar experiment was conducted for a CD-10 unmanned helicopter (Han He Aviation Technology Co. Ltd., Wuxi, China) and a second order regression model for the density of spray droplet deposition was established. The regression model includes the effects of flight altitude and speed, their second order terms, and an interaction term to model spray droplet deposition (Qiu et al. 2013). Spray altitude at 7 and 7-m spray swath were reported to produce the most even droplet distribution for an N-3 unmanned helicopter (Research Institute on Simulation Technique, Nanjing, Jiangsu China) sprayer in a test for maize (Qin et al. 2014). A flight altitude of 2 m and average flight speed of 2–3 m/s produced greater droplet deposition and uniformity than other altitudes (1 and 3 m) and speeds (3–4 and 4–5 m/s) for a HY-B-10L unmanned helicopter (Gao Ke Xing Nong Co. Ltd., Shenzhen, Guandong, China) spraying on rice (Chen et al. 2016). In a spraying test for a QJ-460 quad-rotor UAS (Lv Wei Shi Aerial Plant Protection Technology Co. Ltd., Zhuhai, Guangdong, China) in a citrus orchard, the citrus canopy received the maximum droplet deposition density at a flight altitude of 1 m above that canopy (compared with altitudes of 0.5 or 2 m) (Zhang et al. 2016c). Citrus trees with a central open canopy received greater droplet coverage and density than round-top canopy trees in a test in which both types of trees were sprayed by a 3 W-LWS-Q60S quad-rotor UAS (Crop Guardian Aerial Plant Protection Co., Zhuhai, China) (Zhang et al. 2016b). For studying spray droplet deposition and drift patterns, general test approaches have been proposed. These include test frameworks for spray droplet deposition at top, bottom, upwind and downwind parts (Wang et al. 2016), the NJS-1 type open-circuit, closed low-speed wind tunnel (Ding et al. 2015), and NERCIEA-I (Tang et al. 2016) and NERCIEA-2 high-speed wind tunnels (The Technology Innovation Strategic Alliance for Agricultural Aviation, TISAAA 2017).

Table 2 Spray droplet distribution test results from the literature for agricultural UAS sprayers in China
Full size table

UAS positioning and navigation

A remote control UAS test system based on the Laboratory Virtual Instrument Engineering Workbench was developed to track real-time UAS flight routes (LabVIEW, National Instruments, Austin, Texas USA) (Zhang et al. 2015). This route tracking system was assembled in a UAS flight real-time monitoring system. The monitoring system was able to track the UAS flight route in real-time and can be used to assess the performance of UAS operators or auto-navigation systems (Zhang et al. 2016a). Precise flight route control is essential, especially with poor Global Navigation Satellite System (GNSS) signal reception. Positioning methods other than GNSS can ensure a proper UAS flight in such a situation (Table 3). For example, a positioning reference model was combined with GNSS positioning in the controller for rotary-type UAS. Performance of the controller was tested, and strong robustness and tracking ability were reported (Zhao et al. 2016b). A mixed positioning system with a GNSS device and General Packet Radio Service (GPRS) was tested, showing positioning improvement over a positioning system with GNSS alone (Yuan 2016). A Lite V2 LIDAR laser range finder (PulsedLight Inc., Bend, OR, USA) and an MPU6000 accelerometer/gyroscope (InvenSense Inc., San Jose, CA, USA) were used to control altitude for a quad-rotor UAS. This system could control altitude deviation of the UAS within 0.1 m from a pre-defined flight altitude while the deviation for a GNSS device and pressure gauge control system was 0.5 m (Zhao et al. 2016a).

Table 3 Agricultural UAS positioning and navigation systems for plant protection from the literature in China
Full size table

UAS auto navigation is an important component for precise aerial spraying (Table 3). Qu and Liang (2008) developed a real-time wind field estimation and trajectory correction navigation system with a dead reckoning system/radio positioning system/GNSS. Chen and Lu (2012) developed a navigation system based on a TMS320F2812 digital signal processor chip (Texas Instruments, Inc., Dallas, Texas USA) that incorporated GNSS positioning, infrared sensor and electronic compass functions, and was extended with asynchronous communication. Liu and Wang (2013) designed a strap-down inertial navigation system aided by an airborne data collection system and Baidou navigation satellite positioning. Zhang et al. (2016d) presented an obstacle avoidance control system for a low-altitude and slow-speed unmanned helicopter, based on an improved artificial potential field approach. That system substantially reduced avoidance path length and avoidance time for short obstacles and high square-shaped obstacles. Considerable spray uniformity and flight route precision were achieved by an auto-navigation and auto spraying system developed for an N-3 unmanned helicopter in an experiment (Xue et al. 2016). In the future, computer mapping technology could provide digital maps of individual farms in a village for a UAS using an auto-navigation function, which would increase the potential and precision of aerial applications for farm production.

Variable-rate spray technology

Variable-rate spray technology is the core of very precise and efficient UAS aerial spraying. Nozzle designs were developed and tested to achieve variable-rate spraying (Table 4). An electrostatic nozzle design reduced the use of pesticide by 5.22 L ha−1 while improving the control rate for a pine tree moth, Dendrolimus spp., by 33.8% (Zhou et al. 2012). In a study of UAS spray control of wheat midge, Sitodiplosis mosellana (Géhin), two types of nozzle atomizer were tested, and a centrifugal atomizer was preferred over a hydraulic atomizer (Gao et al. 2013). A rotation centrifugal nozzle was preferable to a pressure nozzle in a citrus spraying test of the QJ-460 quad-rotor UAS (Zhang et al. 2016c). Combinations of different spray motor voltages, spray pressures and nozzle diameters were tested for effective use of a rotary hydraulic atomizing nozzle spray system, and optimal sets were reported (Ru et al. 2014). Wen et al. (2016) presented an ultra-low-volume swirl nozzle design whose atomization characteristics met the requirements of ultra-low-volume spraying. Modification of the nozzle design achieved adjustable spray amount and droplet size. It could be applied to variable rate spraying.

Table 4 Summary of variable-rate aerial spraying for agricultural UAS in China from the literature
Full size table

Aerial application of other crop management materials

In addition to aerial spraying of chemical pesticides, UAS users in China are exploring aerial application of other crop management materials (Table 5). For example, Li et al. (2013) reported the practical use of UAS for sugarcane pest control, releasing a bio-control agent, Trichogramma chilonis Ishii, via its infested host egg solution. UAS can also be used as a tool for supplementary pollination. Wang et al. (2013) suggested that a working altitude of 7 m in downwind direction flight minimized the impact of a tested UAS (Z3 helicopter by Gao Ke Xing Nong Co. Ltd., Shenzhen, Guangdong, China), after they explored optimal parameters for the unmanned gasoline-engine single-rotor helicopter in supplementary pollination of cultivated rice. Li et al. (2014) conducted similar experiments for a uniaxial, single-rotor electric-powered unmanned helicopter, reporting optimal working parameters at a speed of 1.56 m/s, payload of 14.05 kg, and flight altitude of 1.93 m. Seed broadcasting via a 12-axis rotor wing UAS and an artificial seed broadcaster was compared, and the results suggested that UAS seed broadcasting is feasible (Li et al. 2016).

Table 5 Summary of agricultural UAS applications of other crop management materials in China from the literature
Full size table

Imaging and remote sensing

Vegetation indices extracted from UAS-acquired images have been used for vegetation status monitoring and disease and pest detection in China (Table 6). Han et al. (2005) included UAS-acquired images for growth status estimation of winter wheat, and this was suggested for practical use. Zhu et al. (2010) combined UAS-acquired imagery with other imagery to conduct nitrogen diagnoses for rice leaves and canopy. Duan et al. (2014) reported that the leaf area index (LAI) of three crops (maize, potato and sunflower) from 128 bands and 0.7-m resolution UAS-acquired hyperspectral imagery could be estimated with reasonable accuracy (13.6% relative root mean square error relative to in situ LAI measurement). High correlations between the wheat powdery mildew disease index and vegetation information have been reported (Qiao et al. 2006). The vegetation information was extracted from digital images acquired by a G2 digital camera (Canon Inc., Tokyo, Japan) onboard a gas-powered UAS or with a FieldSpec handheld spectroradiometer (Analytical Spectral Devices, Inc., Boulder, Colorado, USA). The authors proposed their use for detecting and monitoring wheat disease. Hu et al. (2013) compared three algorithms with images obtained by a double-spectrum camera onboard a UAS, reporting higher accuracy of diseased pine tree detection using a Weighted Support Vector Data Description algorithm.

Table 6 Summary applications of UAS imaging and remote sensing for plant protection in China from the literature
Full size table

Although UAS remotely sensed imagery has been increasingly used for vegetation monitoring, it has great potential in pest detection and identification for UAS aerial applications. UAS remote sensing has an advantage over satellite remote sensing at small scale and site-specific crop management, because it is more flexible to schedule and can acquire high-resolution imagery data under favorable atmospheric conditions (Lan et al. 2010) using a variety of image sensors. This is essential when time series of good-quality data are required for vegetation monitoring. With technological advances in big data, image processing and deep learning, UAS imagery acquired by digital cameras or hyperspectral sensors could promote UAS aerial applications to pest detection and pest identification.

Challenges and prospects

Agricultural aviation is growing fast in China. However, compared with other parts of the world with large-scale agricultural production, this growth lags behind the needs of agriculture in the country. The number of large agricultural aircraft is only ~0.13% of those worldwide, and covers 1.7% of arable land in the country (Zhou et al. 2013). Given increasing labour shortages and concerns for environmental health, UAS have received much attention in China. In July 2016, there were 108 agricultural UAS manufacturers and 27 UAS research institutes nationwide (TISAAA 2017). Among the manufacturers, 80 (59.3%) provide UAS with both imaging and aerial applications, 49 (36.3%) focus on UAS aerial application alone, and 6 (4.4%) are dedicated to imaging.

There are concerns that must be addressed for the growth of agricultural UAS in China. Factors that slow the development of agricultural aviation in the country were summarised by Luo (2014). These factors are the needs of administrative policies, regulations, strategies and standards, research on supporting technologies, trained professionals, and recognition and supporting services from the general public. Many of these factors also apply to agricultural UAS. For example, a lag of applicable administrative regulations behind the agricultural UAS industry is a major concern for agricultural UAS in China. Security issues also stress the urgent need of regulations. The Civil Aviation Administration of China has recently released two provisional regulations. One for regulation for civil UAS pilots was released in November 2013 (The Civil Aviation Administration of China, CAAA 2013), and the other for civil uses of small and micro UAS was released in December 2015 (CAAA 2015). The first regulation listed qualifications for UAS pilots and flight altitudes and ranges for four UAS categories—micro, mini, small and large. UAS for plant protection purposes were listed as a separate category in the second regulation. Relevant qualifications and restrictions that are less strict than other categories for UAS heavier than 7 kg were also specified. Another regulation on business use of UAS was issued as a draft for review in November 2015, and may be officially enforced in the near future.

Cost is a factor that has limited intensive use of UAS for plant protection in China, though agricultural UAS have been successfully used in Asian countries with small farm sizes, such as Japan and Korea (Zhou et al. 2013). China and four other developing Asian countries (India, Indonesia, Bangladesh and Vietnam), have the largest number of small farms in the world (Thapa 2009). Farmers could greatly benefit from aerial applications in agriculture. However, as with many technologies, the high cost of UAS deters their application in developing countries. In China, although increasing demand promotes the agricultural UAS industry, agricultural UAS applications remain in need of improvement in many aspects. Currently, an average domestic agricultural UAS in China generally costs in excess of 100,000 CNY (USD 14815). Prices of imported models are higher. Before technology advances enable UAS to be affordable to the general public, governmental subsidies and aerial application service providers can facilitate UAS agricultural use. In several provinces (e.g., Henan), a grower could spend as little as a third of the regular price of a UAS after subsidies. Aerial application companies provide UAS spraying services with a price range from CNY 100 to 200 (USD 15–30) ha−1. This service could improve efficiency by a factor of 30 (Liu et al. 2014) while reducing production costs by CNY 105 (USD 15) ha−1 (Zhou et al. 2013). In addition to UAS production with larger payloads and longer flight ranges at acceptable cost, the development of a reliable auto navigation system, very effective adjustable spray nozzles and related spraying methods for such UAS are necessary for their widespread use.

Similar to the need for regulations, the lack of an industry standard or protocol is another factor limiting widespread agricultural use of UAS, at least from the aspects of portability and governmental subsidies. Because of the lack of a mutually agreed industry standard or protocol, there are many types of agricultural UAS in the Chinese market, e.g., gas- or battery-powered unmanned helicopters, fixed-wing aircraft and rotary-wing aircraft of various sizes. The varieties of UAS types and sizes diminish their portable uses and compatibility. Such varieties also make governmental subsidy policy difficult to implement and enforce. This could deter crop growers from buying a UAS because of not receiving governmental subsidies, for the reason that UAS specifications have not been agreed to by those specifying subsidy policy.

An ultimate UAS application scenario in agriculture would be a combination of aerial remote sensing, real-time image processing and variable-rate aerial application technology as described by Lan et al. (2010), i.e., aerial imagery processed instantly to retrieve and diagnose crops, soil, pest and damage information, then variable-rate spraying used accordingly. This could be described as data gathering, data processing and spraying. Initial pest detection and identification are critical to the full scenario. Image processing methods (see Guan et al. 2009; Sankaran et al. 2010 for review) could be used for detection of disease, pests and weeds. Sankaran et al. (2010) suggested that some disease detection methods could be incorporated into autonomous agricultural vehicles or robots for real-time plant disease detection. These methods include the aforementioned remote sensing and digital photos. Other imaging sources, such as fluorescence spectroscopy and fluorescence imaging, could also be used by UAS and processed for plant disease diagnosis. The image processing detection stage could be facilitated by a fast data link between UAS and ground stations, to use computation power. Big data and deep learning approaches could also be included in image processing for pest detection and identification stages, as done for 14 crop species and 26 diseases in Mohanty et al. (2016) or for coffee rust development rate in Calvo-Valverde et al. (2016). In another method, an electronic nose system could be installed on a UAS for sampling or detection, once the volatile organic compound profile for a specific plant disease has been well established. Aerobiological sampling and real-time analysis may be another approach to disease detection.

However, cost will continue to be a key factor limiting widespread use of UAS in plant protection for individual farmers in China, unless the operational farm size increases. Even if cost declines to reasonable levels, a practical issue is frequent takeoff and landing for reloading application material, changing batteries or refuelling for small and light UAS. The majority of UAS in China are battery-powered and their payload capacities are less than 15 kg (Fig. 2). For example, for a battery-powered UAS with a typical 15–20 min flight duration, it is suggested to change its battery after each flight task. However, the ultimate UAS application scheme for plant production described above will increase in cost with technology advances, possibly requiring additional costs to users. In addition to operating a UAS per se, pest surveys, flight route planning and variable-rate spray settings based on pest survey results require additional work and expertise. These UAS operations and related pest work are also extra burdens for crop growers, and should be left to specialized operators. A solution is to use UAS as an on-demand service provided by local governments, agricultural co-operatives, and specialized companies for crop growers. The service could be a chain including UAS application and farming material companies, as proposed in Leng et al. (2017). Sales and logistics companies could also be incorporated in the chain.

Conclusions

The agricultural use of UAS in China will increase in the near future with support from new policies on agricultural aviation, land transfer and large-scale farming. Crop growers and the agricultural industry could benefit from its high efficiency. However, agricultural UAS manufacturers and providers must reduce their costs and face the challenges of producing sound solutions for auto navigation, pest detection and identification and variable-rate spraying systems. UAS on-demand application services provided by local governments, agricultural co-operatives and specialized companies may be a practical solution for crop growers in China.