Precision Agriculture (PA) uses traditional farming practices along with technology to make output production more efficient [1
]. PA plays an important role in today’s agriculture production as it can be used to monitor and control the spread of diseases [2
]. For farmers, diseases can be devastating, greatly affecting the output of a crop and limiting the yield produced. Ideally, most farmers would strive to prevent any diseases from occurring altogether, but this is not always the case. Controlling the likelihood of diseases occurring requires advanced knowledge on factors such as vegetation type, soil, and weather, with the latter being the most unpredictable. The timing of disease manifesting is unknown, and it depends on the conditions which are necessary for its development. Events such as rain or high winds can easily transfer diseases between plants. Therefore, once a disease has been identified, measures need to be taken to prevent a serious outbreak which can have a major impact on the yield and quality of harvest for that season and future seasons.
Vineyards are one area in agriculture where precise monitoring can alleviate the damage that could be caused to plants. In addition to diseases, farmers need to be on the lookout for pests such as birds, insects, and other animals which can cause damage to the plants in a vineyard. In an attempt to minimize the amount of damage that occurs, all sections in the vineyard are checked, which is referred to as scouting. Ideally, scouting is to take place at least once a week with special attention given to areas that are prone to diseases and pests. Unfortunately, scouting is known to be a very manual process, requiring a key eye to detail to accurately detect problem spots in the field. It is also a very time-consuming process requiring a large number of hours for a large field to be checked thoroughly [4
]. Since most diseases prefer cool shady areas, such as the underside of leaves, they are often missed by those who are inexperienced [5
To help manage those who specialize in scouting fields for areas that require special attention, one promising approach is the use of Wireless Sensor Networks (WSNs). By using a WSN, sensor nodes can be placed in areas where the probability of a certain event to occur is high. If an event does occur, a notification could be issued to the farmers to indicate where an action to solve the problem should take place. Using this type of monitoring system would be greatly beneficial for those with large fields as micro-climates can often cause minor changes in weather conditions over little distances [6
]. At the same time, these technologies would be even more useful in the cases of heterogeneous environments, like hilly terrains or similar. On top of a WSN, Unmanned Aerial Vehicles (UAVs), commonly known as drones, can also be used to provide real-time data along with high-quality pictures that can help with the detection of the disease. In recent years, due to the rapid technological advancements, commercial use of UAVs has become more common [7
]. With technology now more openly available for use and prices in the range for everyday consumers, custom drone solutions can be designed for almost any type of application. Having the ability to be programmable and function far distances without any human interaction, drones are very flexible devices that can enhance monitoring in the field [9
]. Due to its small stature, its maneuverability, and the ability to access and view areas that a human would not be able to [10
], drones can provide many advantages for PA. One instance where drones have a clear advantage is in taking aerial views. For a human to take aerial pictures, some possible options are through the use of satellites or by flying with a plane, both of which are expensive and not very accurate. By being able to monitor from not only ground levels but also from above would allow farmers to gain a new perspective on the challenges presented to them. Using a drone is a simple and cost-effective solution that can achieve high precision.
IoT devices have several advantages and challenges when it comes to agriculture and viticulture [11
]. In this work, a low-cost hardware implementation for real-time data acquisition and processing in a vineyard is presented. Wireless nodes consisting of small and low-cost Internet of Things (IoT) devices are used for monitoring the soil moisture and the soil temperature. If the monitoring data exceed some predefined values, a drone flies over the area towards the node that reported these data and takes pictures. Then, the drone returns to the base station and forwards the pictures to a control room for further processing.
The novelty of this work is on the introduced system framework and system implementation. Commercially available, low-cost IoT devices were used to build wireless monitoring nodes. A framework is also designed for data transmission. To minimize duplicate messages, a routing protocol was implemented in each node. The energy consumption of each type of node, monitoring and relay node, was measured in and was minimized. Three wireless technologies were examined and compared in terms of energy consumption. A drone was used to locate specific nodes in the field and the localization accuracy was also measured. Finally, two feasibility tests were conducted to examine the performance of the proposed system.
The rest of this paper is organized as follows: in Section 2
, some background information on vineyards and the diseases that can occur are presented. In Section 3
, the related work is reviewed. Section 4
describes the architecture of the proposed system followed by a description of the different hardware components in Section 5
. Section 6
presents results from experimentation with the system, and Section 7
concludes this work.
2. Background Information
To build an efficient monitoring system for a vineyard, it is important to understand the many problems that can occur when operating and maintaining a vineyard. Not only does the perfect amount of water need to be given to each of the plants, farmers consistently need to check for diseases and pests that could severely impact a season’s yield. Diseases and pests have favorable conditions for their development, and these conditions change depending on the disease or pest. In Eastern Canada, the major grape diseases present include downy mildew, powdery mildew, gray mold, anthracnose, black rot, and crown gall [12
]. Each of these diseases has certain conditions that promote the growth and spread of the disease, with the largest factor being the weather conditions in the vineyard. Monitoring the weather conditions is critical to determining if an outbreak of disease will occur and how large of an impact it will make. A summary of favorable weather conditions for each major disease in Eastern Canada can be seen in Table 1
Pests also pose a serious threat to a plant’s health in the vineyard. Each of the individual pests can target different parts of the plant and become a threat. During certain times of the year and certain weather conditions, a pest is more likely to occur and special care needs to be made. A summary of the major pests in vineyards in Northeastern USA can be seen in Table 2
It is important to note the behavioral patterns of the pests. For example, grape leafhoppers do not affect the quality of the grapes significantly when present in moderate numbers, but they can have a rapid population increase during hot and dry years, causing significant damage [13
]. At the same time, heavy rain can increase the chances of downy mildew. It is important to monitor the vineyard for weather patterns that favor particular pests that can affect the yield and health of the plants.
Another effect the weather has on vineyards that can reduce the yield or quality is winter injury. Winter injury causes damage to the tissues inside the wood and buds of the grapevine due to temperatures dropping below a critical level for the species of grape or due to large jumps in temperature over a short period of time [14
]. One of the methods to reduce winter injury in grapevines is to bury the vines or trunks of the vine to protect the plant from the cold temperatures. This is done to ensure that the vine is healthy and to reduce the damage caused by possible disease and pests. With the proposed IoT- based system, when the low temperatures are reported, proper actions can be performed on time, to avoid further damage.
3. Related Work
Due to the major impact a disease or pest can have on the yield of a crop, numerous systems have been proposed and tested in literature in an attempt to halt the spread of the pathogens [15
]. Since preventing a disease from appearing is impossible due to the uncontrollable conditions in the environment, the focus has been placed on early detection of a disease before it manifests into an epidemic. In an attempt to counteract this growing problem, systems that both utilize drones and do not utilize drones have been developed [16
Latouche et al. [17
] were able to determine that if a plant was infected with downy mildew, infected leaves would fluoresce violet-blue under an Ultra-Violet (UV) light. Tests performed showed that they were able to design a portable sensor that was able to successfully detect the presence of downy mildew in a plant in one day after the point of infection. The sensor was completely non-invasive causing no harm to the plant while providing results immediately. Systems that have focused on utilizing drones have had great success in determining the symptoms of diseases. Di Gennaro et al. [18
] used a drone equipped with a camera to take multi-spectral images. The images could then be used to determine several vegetation parameters of the field. The use of the images was able to determine vines showing symptoms of esca disease that human scouts in the field missed. In addition, previous seasons’ weather conditions were used in estimating the likelihood of a disease occurring. While able to detect early stages of the disease, images produced by the drone were not able to confirm if a disease was present in the plant. Another drone-based system for agricultural use is presented in [19
]. They designed a system composed of two sensors, a stereo vision and thermal imaging camera which are attached to a drone. Images taken by the drone were processed using a deep neural network that was able to classify poor crops from healthy ones. When tested in a vineyard, the system was correctly able to detect diseases such as downy mildew, powdery mildew, and acid rot, but only when the occurrence was above a threshold in the monitoring area. Testing also showed that the system was unable to function if the lighting conditions were not optimal for the sensors.
There are also systems that use sensors for smart farming and viticulture. Several aspects of sustainable IoT devices, including the wireless communication, sensing, and systems are presented in [20
]. In [21
], a low-cost wireless system for agrochemical dosage reduction in precision farming was introduced, while in [22
] a low-cost wireless monitoring and decision support system for water saving in agriculture was proposed. A comprehensive review of computer vision, image processing, and machine learning techniques in viticulture is presented in [23
]. In [24
], a wireless underground sensor networks (WUSA) for autonomous precision agriculture (PA) is proposed. The system gathers soil information, from a WUSN in real-time to automatically control the center pivot (CP) irrigation system for precision irrigation.
In comparison with the works in the literature, the novelty of this work lies in the framework and the system implementation. The introduced approach uses commercially available components, design a real-time monitoring system for vineyards, properly program and configure the different components and finally examine the feasibility of the system through real experimentation.
In this paper, an IoT-based system for real-time monitoring in a vineyard by utilizing drones was presented. The introduced system consists of a low-cost IoT network that can be used in the monitoring of plants in a vineyard. When the monitoring values exceed some threshold, a drone is used to obtain photographs of the problematic area which will be sent to the system administrator. The focus of this work was on a low-cost system implementation for real-time data acquisition and monitoring.
Further experimentation is necessary to examine the performance of the introduced system, while cybersecurity issues should also be considered. The experiments presented in this paper are promising for the feasibility of the proposed system. The system was able to report the data in real-time and also locate the node which collected the data that exceed a predefined threshold, with acceptable accuracy. However, proper sensor calibration and parameter configuration are necessary for every different application scenario.