You are on page 1of 5

A Drone-based IoT Approach to

Agriculture Automation and Increase Farm


Yield
Selvakumar Manickam
National Advanced IPv6 Center Universiti Sains Malaysia, Penang, Malaysia.
selva@usm.my

Abstract— This paper proposes a smart agriculture system which problem are can this implementation aplied to other places as
include of soil sampling, drone seeding, pesticide/fertilizers spraying well? As the willingness of citizen is a must consider factor.
and farm condition monitoring through the application of IoT sensors
and drones for new farms and farms practicing crop rotation. Drone Hence, this paper will propose an idea of smart
will collect farm information such as soil humidity and pH value from agriculture system which include of action that farmer can
ground sensor where farmer pre-plug in the farm. The connection perform during pre-farming stages and farming stages.
between ground sensor and drone are using Zigbee due to its capability
of wide range connection. The collected data will then transmit to II. BACKGROUND
gateway through LoRa and the information would then store in the
cloud as data storage and data analysis. Through the proposed
application we believe that the farming work is more automated and Unmanned Aerial Vehicles (UAV) or drone had become
crop yield rate will improve. cheaper in the market, this is because many control
functions can be implemented in software rather than
Keywords—Internet of Thing, Agriculture, Farm having to depend on expensive hardware [5][6].
There is various factor that can affect the crop yield rate
I. INTRODUCTION and it is hard and costly to solve in the past. With the
Food is the essential need for human being, and the ultimate emerging in drone and sensor technology as well as the
source to obtain food is through Agriculture. Agriculture is the network communication. Those issue shall mitigate now.
process of producing food, feed, fiber and many other desired Smart farming and precision agriculture involve the
products by the cultivation of certain plants and the raising of integration of advanced technologies into existing farming
domesticated animals (livestock). [1] practices in order to increase production efficiency and the
In Agriculture, crop yield is the term we defined its output. quality of agricultural products. As an added benefit, they
It is an important indicator to measure the yield of a crop per also improve the quality of life for farm workers by
unit area of land cultivation, and the seed generation of the plant reducing heavy labor and tedious tasks.
itself. Figure 1 shows the long-term wheat yields in Europe [2].
A. Bird Attack
Various paper and news shown that bird attack had
causing serious trouble to the farm. [7][8][9]. The issue not
only occurs now but it happened in ancient time as well. In
the past, people are using scarecrow to make bird scare. But
once the bird realized that it is static and post no harm, then
tragedy will happen to the farm. An alternate to tackle the
issue, farmer trained a falcon to scare away the bird, but the
training process is long and costly.[10][11]
One of the solutions from this paper is that [12], they
suggested to use a sudden loud sound, produced by propane
gas guns. But these solutions, is producing a constant noise
pollution.
Figure 1: Long-Term Wheat Yield In Europe [2]
Another article suggested to detect pest birds through
The mankind population had increased drastically from image recognition and generate random sound to scare the
3.032 billion to 7.53 billion from year of 1960 to 2017. [3] This birds away. [13]. In this article, stational speakers are used
increased amount of population imposed to the statistic in and it is not scalable. Multiple station would have to be
Figure 1. That said, the output of Agriculture directly installed in a large farm.
proportional to human population.
Thus, a drone with ultrasonic sensor and buzzer would
To support the skyrocketing population, an efficient of farm be the great solution to implement. Drones are mobile unit
management need to be study and implement. In this paper [4], so they could cover wide zone. A buzzer and ultrasonic
they suggested to perform rooftop farming. They scope on these sensor are mount to the drone. Through ultrasonic sensor it
two area Nagdhungha and Birauta and their result shows that will detect when there is wildlife close to our farm. Once
the stormwater management cost can be reduced, but the wildlife is detected around the crops, the buzzer will start
beeping to scare animals away from crops. In addition to
the sound produced, birds are also scared the unknown
flying object approaching them as well.

B. Efficiency in Soil Sampling


On large farms, there are segmentation of zones with
different soil properties. A soil map with measured properties
is one of the important information for fertilizer prescription for
the farm and deciding the crop type in next farming season. The
common practice in analyze soil properties is to sample the soil Figure 2: Smart Agriculture Block Diagram
for analysis in laboratory. The analysis is charged per sample
and could be very expensive if sampled per square meter. So, it
is important to collect minimum amount of soil samples that is B. Pre-farming (Soil Sampling)
enough for analysis. The pH value is the key parameter when deciding which
seed need to plant, and the suitable fertilization formula to
A drone could be used to capture the soil color mapping to use. But given the large size of the farm, it would be
form farm management zones [14]. Soil samples could be overwhelmed and costly for the farmer to do decision.
collected from each farm management for analysis.
Thus, soil sampling with aid of drone is necessary. A
drone will take an aerial image and through image
C. Seeding, Fertilizing and Pesticides Spraying Effectivity processing we can obtain soil map color. This aim of the
Seeding effectivity is another major concern to farmer. In process is to break the big farm into sub region. Once we
the finding of this paper [15], it would take five hours to can get the info in sub region, the farmer can send the soil
walking an 80-acre soybean field. In regards of this, our paper sample to laboratory for precise analysis. Figure 3 shown
proposed to use a drone to plant seed in the farm. Similarly, the process of soil sampling.
fertilizer and pesticides spraying could be done in much shorter
time using drones, eliminating the needs of human labor for
such tasks.
In addition to reducing farming workload, through
monitored plants health, fertilizer can be adjusted to optimum
amount, reducing fertilizer cost.
Figure 3: Soil Sampling
III. PROPOSED SYSTEM
C. Pre-farming (Drone Seeding)
A. System Overview As per mentioned in the problem statement, to plant a
There are two phases in our proposed system, phase one, seed in the farm is another tedious work to performed. Thus,
which is pre-farming stage. In pre-farming stage, land tillage is a drone seeding is needed to perform the action. In drone
performed at first. Then, bare soil could be seen from aerial seeding module, our proposed approach is that the drone
vision. The drone will then take an aerial image of the farm carries various type of seed, with the aid of 3D mapping our
location. This give an overview to farmer, based on the farmer can planned a pre-planned route based on the input
processed image, farmer know the actual size of the farm, and from soil sampling. Thus, the seed will plant in the correct
thus the route of drone seeding can be planned accordingly. region. Figure 4 shows the process mapping and Figure 5
Another service can provide in the pre-farming stage is soil shows the image illustration of drone seeding.
sampling. In this paper [14], they had demonstrated a protype
on the way to gather information from soil through aerial image
and augmented reality software. But their implement is using
image processing with AR technology which we don’t need in
our implementation. Thus, in our design we removed the AR
section and include the IoT capability.
In phase two, the drone is capable of gather the information
of the overall farm status from its sensors and ground based IoT Figure 4: Seeding Figure 5: Seeding Illustration
sensors and send the information to cloud service for data
storage and analytic. The information includes, the soil
humidity, pH value, farm temperature and wildlife. D. Farm Monitoring
Another service will implement in phase two are, drone self- Farm monitoring are the core of our approach. In this
charging module. In this module, a capacitor in the drone will module, drone able to collect sensors data consisting of
detect the battery level, once the battery percentage is below the humidity of soil and temperature of farm. From the data
certain threshold amount, drone is capable to fly to a base collected, optimum amount of fertilizer and water is
station for self-charging. calculated and supplied. Last, it needs to constantly
feedback NVDI images to cloud to perform normalized
For the communication protocol between ground difference vegetation index (NDVI) calculation. NDVI is
sensor and drone we proposed to use Zigbee, while the an algorithm used to determine crop health and growth
communication between drone and gateway are LoRa. monitoring. Healthy vegetation (chlorophyll) reflects more
Figure 2 demonstrate the block diagram of our proposed near-infrared (NIR) and green light compared to other
system. wavelengths. But it absorbs more red and blue light. This is
why our eyes see vegetation as the color green [16].
Equation 1 below shows the formula of NDVI.
NDVI = (NIR – Red) / (NIR + Red) ----------Equation (1)
IV. EQUIPMENT
In short, the higher the NDVI the greater the health of plant.
Low NDVI means low vegetation. Figure 6 shows the The basic equipment where needed for the proposed system
comparison of Real Color Image and NDVI Image. Thus, based are outlined in Table 1.
on the info return from NDVI, farmer can know which area
require “special treatment” for the fertilizer. Fertilizer amount
could be reduced at health area and more fertilizer could put on TABLE 1: EQUIPMENT NEEDED
less healthy area. Humidity sensor are used to determine the Current Sensor Monitor and optimize power drain,
moisture content. In our implementation, a humidity sensor safe charging of internal batteries.
would plug in the soil, they collect and transmit data to drone.
Detect fault condition within drone.
Ultrasonic sensor in this module are used to detect if any
bird close to farm, if they are close then the buzzer will “buzz” Humidity Sensor Plug in soil to detect soil humidity
to chase out the bird. Additionally, the data will transmit to level
farmer connected device to bring their awareness.
Temperature Provide overview of farm temperature
Sensor level.
Drone Provide GPS location, maintain flight
level.
HD Camera Provide Image for soil sampling
(1080p) analysis.
Figure 6: Comparison of Real Color Image and NDVI Image Servo Perform various function.
Spraying fertilizer.
E. Fertilizer and Pesticides Spraying Buzzer Produce beeping sound.
With the IoT solution, suitable fertilizer amount for
each farming zone are calculated. NVDI Sensor Collect NVDI wavelength light for
crop health assessment.
The task of spraying of fertilizer and pesticides are
scheduled by software, without need of human intervention. For the camera, we opted for 1080p as it provided best
At scheduled time, software set the drone to spray fertilizer feasibility in all class. 720p appear to be less detail while
and pesticides at the planned route. Aerial spraying by drones the transmission rate required for 4k is not feasible for the
has high efficiency, it could cover much bigger area than implementation.
done by human labor. By using drones, farmers avoid the risk
of heat stroke and poisoning. V. COMMUNICATION PROTOCOL
For the communication between ground sensor and
drone, we had done various research and finding are
F. Farming(Drone Charging) tabulate in Table 2. Per our finding, there is no specific
An autonomous charging solution are also proposed. A network can be used, it is purely depending on the size of
current sensor in drone can detect the battery level of a drone, the farm. If the farm size is smaller, the drone can fly lower
if the battery is fall under threshold percentage, then it will fly to collect the data as the range of BLE are shorter. If the
to a charging pad for charging. Since the charge panel is placed farm size is larger, then we recommended to use Zigbee as
outdoor, thus the charging pad should be water resistance it can provide up to 50m transmission rate.
(IP55). But in our implementation, we are using Zigbee as our
The charging data will transmit to cloud for storage and communication protocol. Zigbee uses 2.4GHz frequency
further analysis. The user can manage and monitor the data via to transmit low packet data. Moreover, it also shows a
API and graphical user interface (GUI). User can monitor the characteristic of larger transmission range over BLE.
heath of the battery and grant awareness if a battery
replacement needed. Figure 7 show an example of drone and
charging pad.

Figure 7: Drone and Charing Pad


TABLE 2: NETWORK COMPARISON 5 hours to walk but with the aid of drone, this would save a
lot of effort.
While in the farm monitoring stages, through IoT approach
we can easily collect and centralize the data and do analysis
on temperature, humidity and pH of farm. The correct
amount of fertilizer and water could spray to the farm.
Moreover, through the data analysis, farmer predict what
might happen in future and prepare solution for that.
Another great feature in our approach are self-charging, the
current sensor in the drone are able to detect if the battery
has lower than the threshold level, if it falls beyond that, the
drone will automatically fly to a charging station.

VII. FUTURE WORK


Most BLE devices communicate with each other using a simple In our approach, we had proposed a complete system
point-to-point network topology enabling one-to-one device to improve the efficiency of Agriculture through the aid of
communications. For example, a smartphone can connect to drone and IoT. The process includes of pre-planning and
multiple device such as speaker, smartwatch, heart rate monitor planning stage. The approach shall provide one-end
and others but these devices cannot communicate among each solution to farmer - from planning to harvesting. But the
other. While in contrast, a mesh network has a many-to-many adoption rate is one of the major issues that we could
topology. Each device would able to communicate with each thought. Farmer are used to the traditional pattern of
other in the mesh. farming is unlikely to adapt to the new system.
Zigbee is an IEEE802.15.4 based specification for suite of Another concern is the cost and complexity of the
high-level communication protocols used to create personal system. The system is a bundle of multiple applications
area networks for data collection. A typical Zigbee mesh and require a few resources such as drones, cloud access,
network only require one coordinator is needed form Zigbee and gateway infrastructure. Small farm owners are
network, the coordinator stores all the critical information about unlikely to adopt the system due to ROI concern.
the Zigbee network including encryption keys. In future, an extended work that researcher could
Zigbee Tree Network have interconnected Star network looking into are joint the proposed system into blockchain.
operate in Beacon mode. The Coordinators and Routers Blockchain is able of keeping ledger of many types of
perform frame forwarding, and there are buffer frames among data, for example, purchase order, invoice, banking info
IoT devices. The network is always stay “ON” to listen for and many more.
transmissions from nodes, the coordinator cannot sleep but the Through this development, it would form a
device can sleep. All communications must go through revolutionary ecosystem of agriculture, imagine that a
Coordinator/ Router. system that is capable capture the current need and even
For the communication between drones and ground station, predict future demand.
LoRa protocol is used. LoRa enables communication distance
of more than 10km in rural area. The maximum data transfer
rate is 300kbps. [17] To cater for the limited bandwidth, the VIII. CONCLUSION
large volume data of HD images are indexed and stored on
In this paper, we had discussed the service and the
drone, and only upload and mapped by software when it is
solution that would tackle the traditional farm issue.
docked on station.
Through the implementation, the workload and risk for
farmer shall reduce through the farming process. The
implementation shall show improvement in the crop yield
VI. DISCUSSION rate. Moreover, we believe it would benefit the farmer and
In our approach, we had proposed a smart agriculture society in other perspective, as the knowledge level farmer
system, through this system farmer can increase the crop yield is indirectly improved.
rate. The process includes of pre-planning and planning stage.
To compare our approach with previous work the best way is IX. REFERENCES
to describe with a scenario. [1] Safety and Health in Agriclture. (1999). International Labour
Organization
Consider we have a wheat farm where it is just harvest, and [2] Max Roser and Hannah Ritchie (2019) - "Yields and Land Use in
farmer would like to change the farm to corn farm. In the pre- Agriculture". Published online at OurWorldInData.org. Retrieved from:
farming stage, they need to identify the pH of soil in different 'https://ourworldindata.org/yields-and-land-use-in-agriculture'
sector in the farm and then use suitable fertilizer to provide [3] World Population. (2019-05-12). Retrieved 2019-06-07, from
http://worldpopulationreview.com/continents/world/
enough nutrient to the soil. Traditionally these need to be done
[4] Jha Ritesh Kumar, Bhattarai Natasha, KC Suraj, Shrestha Arjun Kumar,
manually, but with the aid of drone we just need to take a Kadariya Manahar (2019). Rooftop Farming: An Alternative To
picture and perform software analysis the work effort would Conventional Farming For Urban Sustainability. Malaysian Journal of
reduce drastically. Sustainable Agriculture, 3(3): 12-16.

After soil condition is addressed, drone seeding will then [5] C. Zhang, J. M. Kovacs, ”The application of small unmanned aerial
systems for precision agriculture: a review”,
perform respectively. Per finding, 80-acre farm would take
[6] Anderson, C.(2014). Agricultural Drones, 1.
[7] Staff, D.(2019). Agriculture. Birds Impacting Agricultural Crops a Major
Concern, 1.
[8] Abd, El-Aleem & Saad, Soliman & Desoky, Abdelalem. (2015).
DAMAGE CAUSED BY BIRDS AND RODENT IN FIELD CROPS
AND THEIR CONTROL. 2014. 169-170.
[9] Minnock, P. (2017,7 7). Bird Attacks The Biggest Threat To Crops In
Rup-Up To The Harvest Season. Retrieved from
https://www.independent.ie/business/farming/tillage/bird-attacks-
thebiggest-threat-to-crops-in-runup-to-the-harvest-season-
35890225.html
[10] Wang et al. “Psychological warfare in vineyard: Using drones and bird
psychology to control bird damage to wine grapes.” Crop Protection,
2019.
[11] Martin,G. (2015). Contributor. The $250,000 Bird: Falcon Hunting Im
Qatar, 1.
[12] Wright, Ernest N., “BIRD PROBLEMS AND THEIR SOLUTIONS IN
BRITAIN’(1982). Proceedings of the Tenth Vertebrate Pest Converence
(1982).50.
[13] C. Pornpanomchai, M. Homnan, N. Pramuksan and W. Rakyindee,
"Smart Scarecrow," 2011 Third International Conference on Measuring
Technology and Mechatronics Automation, Shangshai, 2011, pp. 294-
297. doi: 10.1109/ICMTMA.2011.644
[14] Janna Huuskonen, Timo Oksanen, Soil sampling with drones and
augmented reality in precision agriculture, Computers and Electronics in
Agriculture,Volume 154,2018, Pages 25-35, ISSN 0168-1699
[15] Gupta, Lav & Jain, Raj & Vaszkun, Gabor. (2015). Survey of Important
Issues in UAV Communication Networks. IEEE Communications
Surveys & Tutorials. 18. 1-1. 10.1109/COMST.2015.2495297.
[16] What is NDVI (Normalized Difference Vegetation Index)? - GIS
Geography. (2019). Retrieved from
https://gisgeography.com/ndvinormalized-difference-vegetation-index/
[17] J. A. Godoy, F. Cabrera, V. Araña, D. Sánchez, I. Alonso and N. Molina,
"A new Approach of V2X Communications for Long Range Applications
in UAVs," 2018 2nd URSI Atlantic Radio Science
Meeting (AT-RASC), Meloneras, 2018, pp. 1-4. doi: 10.23919/URSIAT-
RASC.2018.8471484

You might also like