See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/357394735

HydroIoT: An IoT and Edge Computing based Multi-level Hydroponics System

Preprint · December 2021

CITATIONS READS
0 101

4 authors:

Sanket Salvi Samyak parmeshwar Sawargaonkar

National Institute of Technology Karnataka Vishwakarma Institute Of Information Technology, Pune
30 PUBLICATIONS   162 CITATIONS    2 PUBLICATIONS   0 CITATIONS   

SEE PROFILE SEE PROFILE

Pramod Jain Vishwas Bhaktavatsala

Nitte Meenakshi Institute of Technology Reva University
2 PUBLICATIONS   0 CITATIONS    1 PUBLICATION   0 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Industrial Internet of Things View project

Visible Light Communication View project

All content following this page was uploaded by Sanket Salvi on 27 September 2022.

The user has requested enhancement of the downloaded file.

 HydroIoT: An IoT and Edge Computing based
Multi-level Hydroponics System
Sanket Salvi∗ , Savitha S† , Samyak Sawargaonkar‡ , Vishwas B§ , T K Harshitha ¶ , Pramod Jain S Ak
∗¶k Department of Information Science and Engineering,†§ Department of Computer Science Engineering,
‡ Department of Mechanical Engineering,
∗¶k Nitte Meenakshi Institute of Technology, † Christ University, § REVA University, Bangalore, Karnataka, India
‡ Vishwakarma Institute of Information Technology, Pune, Maharashtra, India

Email: [email protected], † [email protected], ‡ [email protected],

∗
§ [email protected], ¶ [email protected], k [email protected]

Abstract—The depleting area of cultivable lands is increasing

demands for implementing improved techniques that could use
less space and produce more than traditional farming. This
situation is common in all the developing and under developed
countries. With a motivation to contribute towards providing
solution to this growing problem of food scarcity, a Multi-level
Hydroponics System is proposed. The proposed system combines
best of all trending technologies like IoT, Edge Computing and
Computer Vision and applies it to Hydroponics. A cultivation
estimation system based on image processing is implemented
and accuracy of the same is tested with actual produce. The Fig. 1: Generic Hydroponics System [3]
crop used for the proposed system is corn as it serves as best
fodder for cattle. It was observed that with proposed system upto
95% accuracy in estimating fodder produce was achieved. efficient hydroponics. According to Maharashtra’s Department
Index Terms—Hydroponics; Internet of Things; Vertical Farm- of Agriculture, the state is facing a 74% shortage in green
ing; Precision Farming; Edge Computing; Image Processing;
fodder. Hydroponics systems combined with Internet of Thing
(IoT) can be revolutionary in such conditions, the solution can
I. I NTRODUCTION
be implemented with comparatively very less amount of water
With the increase in population, there had been increase at low cost.
in urbanization too, which has led to reduction of cultivable IoT for agriculture [4] had shown promising results by
lands. Smarter and non traditional alternatives to framing are automating the supply of water as per crop’s requirement
now most sought after topics to meet the food requirements using predefined moisture level for different crops. Also by
of the growing population. Soil-less farming is one such leveraging the power of Computer Vision (CV), a detailed
technique that has garnered lot of attention in recent times, analysis of crop diseases and cultivation estimation can be
due to its efficiency and no demand for dedicated agricultural provided [5]. Thus, it shows that an integrated system of IoT,
lands. Hydroponics and Computer Vision can provide more efficient
According to studies, plants which are grown by means of insight to produce better yield.
hydroponics grow faster and are healthier than plants grown The objective of this study is to built a low-cost household
in the soil [1]. This is because, the nutrients are directly setup with minimal man power to provide constant fresh
supplied through the water in the form of aqueous solvent. supply of green fodder for livestock. An hydroponics system
Hydroponics, as shown in Fig.1, can be a major alternative for is proposed with the layers of computer vision and IoT
growing plants other than the traditional way [2]. In an attempt to test automation and quality in fodder production. The
to boost public interest in this area, hydroponic farms were also contents of paper are arranged as follows, where section II
implemented in India under which government established 11 highlights various works that inspired our proposed system
hitech Hydroponics unit under Rashtriya Krishi Vikas Yojana followed by section III describing details of our proposed
(RKVY) at research complex, Goa of the Indian Council of model. Section IV and V provides insight on hardware and
Agriculture Research (ICAR) and Dairy Co-operatives of Goa software requirement, its integration for implementation and
state. The earlier setup which were installed in India were detailed discussion of results. Finally Section VI provides the
not cost efficient and had high cost of initial installation. conclusion of the work with few future line of actions.
Moreover, the technicality for installing was huge learning
curve compared to the traditional farming, thus hydroponics II. R ELATED W ORK
being less popular option in practicality. However, the current Growing a plant without soil helps to increase the yield of
demand is driving need for faster, scalable, mobile and cost crops in short duration for any climatic condition. In paper
 Start

Input from Input from Input from

Input from pH
Input from

Water Level Temperature Humidity

Level Sensor Light Sensor
Sensor Sensor Sensor

Combine data to
form a bundle

Send Data Bundle to

Fig. 2: HydroIoT Block Diagram Raspberry Pi

Capture Image of
[6], a survey about technology and techniques are discussed Crops
to collect a review on automatic monitoring of hydroponics
farms using IoT. In an attempt to optimize resources needed Find Height of the
for hydroponics and make it more efficient paper [7] proposed Crops

application of Machine Learning(ML) techniques. It uses

inputs from sensors and applies KNN and Lasso Regres-
sion algorithm to maximize benefit. These works provided No
Set Cultivable Value to if height >=
required insight on precision farming and hydroponics. These 0 threshold
works guided towards providing clear understanding about the
domain and motivated towards identifying future trends and
Yes
research challenges in hydroponics. Monitoring growth of the
crops at all stages in basic principle of Precision Farming. Estimate cultivable
Thus, a Computer Vision (CV) based system was proposed volume
in paper [8] to estimate plant growth vitals such as height,
leaf length and dry matter. In a similar approach, images Append height value
and cultivable volume
of farm taken from UAV were analyzed to automate the to the bundle and post
counting process. The proposed system in paper [9] [5] uses bundle to cloud
Convolutional Neural Network (CNN) to identify regions of
interest using pre-trained dataset of multi-spectral images of
Display Data on
farm. These works inspired the cultivation estimation module Mobile Application
of our proposed system.
Various automation techniques for hydroponics leveraging
IoT have been proposed recently. Aquaculture integrated hy- Stop
droponics assisted with sensors for Mint plants was proposed
in paper [10]. The vital nutrients required for growth of Fig. 3: Overall HydroIoT Flowchart
crops were supplied, monitored and controlled by using water
level indicator, pH level sensor, digitally controlled valves
connected to Raspberry Pi or Arduino were proposed in [2], into Overall Proposed Workflow, Hydroponics Water Manage-
[11], [12]. An application that provides remote access to the ment Workflow and Cultivation Estimator Workflow.
automated Hydroponics system was proposed in paper [13]– A. Overall Proposed Workflow
[15]. An improvement over existing system was proposed in
paper [16] [17] that could not only provide remote monitoring An overview of various components involved in proposed
capability but also it could plan and predict the harvest while HydroIoT system is shown in Fig.2. The proposed layered
keeping power consumption to minimal. Our proposed system arrangement is based on Edge Computing Architecture [18].
in highly inspired by these works in terms of usage of sensors, The components are arranged four basic functional layers,
actuators and controlling unit. namely as Hydroponics Sensor Module, Cultivation Estima-
tor Module, Cloud Storage and Visualization and Mobile
Application. The first layer will consist of various sensors,
III. P ROPOSED F RAMEWORK
actuators, communication module and controlling unit required
In this section, the proposed HydroIoT system is discussed. to operate the hydroponics locally. Sensors which determine
To facilitate better understanding the section has sub-divided water level, pH, temperature, humidity and light intensity will
 servo will turn off the inflow. This local sub-system will also
Start monitor pH level.

C. Cultivation Estimator Workflow

Capture Fodder Perform Row Wise Find Number of The Cultivation Estimator will be run on local server which
Batch Image sum edges
would a raspberry pi. The camera attached to the server allows
to capture pictures of the fodder batch. A striped straw of
Inverse Crop the Estimated Fodder known height is strategically placed behind the batch. The
Convert RGM
Bounding box from Height, Volume,
image to Grayscale
Binarized Image
number of visible stripes will be used to estimate the average
Weight and Price
height of the fodder batch. Fig.4, shows the sequence of image
manipulation techniques used to obtain the results. At first,
Extract the Largest
Upload values on the image will be converted to grayscale and then binarized to
Binarize the Image Blob and find
Cloud
bounding box retrieve largest blob. This largest blob will be the fodder, thus
and inverse cropping of image will be performed to obtain
binarized image containing only stripes. Following which by
Stop performing columnar sum, number stripes will be determined.
The number of stripes is inversely proportional to height of the
Fig. 4: Cultivation Estimator Flow fodder. Thus, using estimated height, the system will estimated
volume, weight and finally price of each batch.

IV. R EQUIREMENT A NALYSIS AND I MPLEMENTATION

be used to determine respective values. A servo motor will
be used to control the water level. These sensors and actua- This section provides details on the hardware and software
tors will be connected to a micro-controller to support local requirements along with its integration and implementation.
decision making. A bluetooth communication module will be Table.I shows the hardware and software requirements and
used to send data to a local server with higher computational its justification for usage in the proposed system. The circuit
capacity and internet connectivity. At the second layer, the diagram for the Hydropoincs Subsystem controller is as shown
local server aggregates the received data. It also captures in Fig.5. A common VCC and GND from Arduino is provided
pictures of fodder batch and performs image processing to to breadboard to power the sensors and actuators. The Rx
retrieve desired information. The cloud and app layers will be and Tx pins of HC-05 are connected to pin 11 and 12 using
for storage and visualization of observed data. SoftwareSerial command. Data pin of DHT11 is connected
The overall workflow of the proposed HydroIoT system will to pin 2 of arduino and corresponding library is installed to
be as per Fig.3. The system will take inputs from its various read temperature and humidity values. LDR in connected to
sensors and bundle it into single object by the microcontroller. the analog pin A0, the voltage values in the range of 0-255
This bundled data will be then sent to the local server via are converted to Light Intensity in LUX. Th pH Level sensor
bluetooth. The local server will be equipped with camera. It is connected to pin A0, and values are read using its library.
will capture pictures of the fodder batch and estimates the The water level sensor is connected to pin A1, the voltage
height of the fodder. If the height of the fodder is as per the level is converted to level of water in tray in cms. Pin 3 of
desired height for the number of days that it was kept, then the ardiuno is connected to servo motor to control inflow valves.
system will perform cultivation estimator and return estimated A string is formed with all the values separated by comma and
weight and price for the batch. If condition is not satisfied, the each string separated by semicolon. For example, ”;w,p,t,h,l;”,
system will add ”0” at the corresponding index of the bundle. where w,p,t,h,l are values from sensors. The arduino BT is
The complete bundle will be then posted to cloud(Thingspeak paired with Raspberry Pi BT using blueman and python-serial
Cloud [19]) where bundle will be segregated into fields. The tools. The constructed string is then forwarded to second layer
values for corresponding fields will be updated with received i.e., Cultivation Estimator (Local Server).
data. The same data will be visible on a mobile application At Cultivation Estimator,string split command is used to
built using MIT App Inventor [20]. extract values from each sensor and it is stored in respective
variables. As the Cultivation Estimator is equipped with cam-
era as shown in Fig.6, the picture of fodder tray is frequently
B. Hydroponics Water Management Workflow
taken and a separated function is called to estimated height
The water level management and quality check is an im- and weight of fodder from image. If the height condition
portant aspect of hydroponics. Assuming that an overhead is not satisfied, the received data is directly forwarded to
water tank with desired pH is available, the sub-system will Thingspeak cloud with height and weight parameters as 0
keep frequently checking for the water level using water level otherwise, the estimated height and weight is forwarded. The
sensor. If the water level drops below predefined threshold then height of the crop is estimated by performing image processing
the micro-controller will turn the inflow valve coming from the techniques as shown in Fig.4. As each fodder batch is having
tank using servo motor. If the desired threshold is satisfied the a striped straw of length 30cm in the background, the height
 TABLE I: Hardware and Software Requirements
Hardware Requirements
Component Model Usage
It is used measure
Water Level
RKI2350 the water level in
Sensor
Hydroponic Tray
It is used determine
the pH level of water
pH Level
SEN0161 from 0-14. The range Fig. 6: Cultivation Estimator(Edge Node)
Sensor
5.5-6.5 is desirable
for hydroponics.
Temperature To determine tempe-
DHT11
Sensor rature of surrounding. estimation is done by counting the number of visible stripes.
Humidity To determine humidity
DHT11 Here, the estimated height is calculated as a factor of 6 times
Sensor of surrounding.
The change in voltage the difference between total number of stripes on straw and
is proportional to light number of stripes detected as shown in Eq.1, where, he is
Light Intensity falling on sensor. Thus,
Sensor
LDR
it is used to determine estimated height of fodder, ts and ds are total number of stripes
Light intensity where and number of detected stripes.
fodder is grown.
It is used to control the
Servo Motor SG90
water inflow he = (ts − ds ) ∗ 6 (1)
Communication Module
Bluetooth
HC-05 to send and receive data
Module ve = (he ∗ bx ∗ by )/vd (2)
from Local Server.
All the sensors and actuators
are connected to it. It is a The volume of one fodder batch is estimated using Eq.2, where
Hydroponics programmable board. It also
Arduino ve is estimated volume, bx and by is length and breadth of
Subsystem serves as data collection unit
Uno r3
Controller and relays it to Local Server. the tray. As the fodder does not occupy the entire volume, a
It also provides local decision volume divisor vd is used to get closest to practical volume.
making.
It aggregates all the data and
Raspberry Pi
Local Server 3B
adds new data after image we = ve ∗ 0.05 (3)
processing before its sent to
+Camera
cloud.
Software Requirements By trial and error it was observed that 0.05 works as best
Its used to store processed multiplier to estimate fodder weight. As shown in Eq.3, the
Cloud Storage data from Local Server for estimated weight we is product of estimated volume and 0.05.
Thingspeak
and future retrieval and analysis.
Visualization
Cloud
It also provides tools for This estimated weight is in grams. Thus, to estimate the price
visualization and APIs as shown in Eq.4, weight is converted to kgs and current
A cloud based Android market price per kg is multiplied i.e., 1.5 Rs. per kg.
MIT App App building tools for
Mobile Application
Inventor functionally stable and quick
app development. pe = (we /1000) ∗ 1.5 (4)

|wa − |we − wa ||
Accuracy = ∗ 100 (5)
wa

The accuracy of our system is tested using Eq.5, where each

value is average of 10 iterations.
The Cultivation Estimator also works as an edge node. It
sends the collected data over Thingspeak Cloud. A dedicated
channel is created on Thingspeak Cloud with 8 fields, of
which 5 are for each sensor and 3 for estimated height,
weight and price. The values are posted using Thingspeak API
with corresponding WriteAPI Key. A fully functional app is
developed using MIT App Inventor which pulls the data from
Thingspeak and displays the information. Static value thresh-
olds are applied to color code value and aid understanding its
significance. Values in black represent normal expected values
Fig. 5: Hydroponics Sensor Module Circuit while values in red represent unexpected or outside desirable
range.
 TABLE II: HydroIoT Percentage Accuracy
Est. Actual Est. Actual
Visible Volume Acc.
Height Height Weight Weight
Stripes Divisor (%)
(cm) (cm) (gm) (gm)
6 2 24 22.3 480 261.7 16.58
6 3 24 22.3 320 261.7 77.72
6 4 24 22.3 240 261.7 91.70
6 5 24 22.3 192 261.7 73.36
4 2 36 31.4 720 302.1 38.33
4 3 36 31.4 480 302.1 41.11
4 4 36 31.4 360 302.1 80.83
4 5 36 31.4 288 302.1 95.33
(a) Captured Image (b) Processed Image
Fig. 7: Image Processing I/O

(a) Water Level (b) pH Level

(c) Temperature (d) Humidity

(a) Desired Values (b) Undesired Values
Fig. 9: Value Alerts

a continuous supply of water was maintained using values and

(e) Light Intensity (f) Height motors.
The values received from each sensor were logged and
forwarded to Thingspeak Cloud. The results of cultivation es-
timator are as shown in Fig.7 where Fig.7a shows the original
captured image after 7 days of germination. Fig.7b shows the
processed image. After removal of the largest blob, the number
(g) Weight (h) Price of stripes from remainder image was extracted. Table.II shows
the accuracy for various combinations of Volume Divisor. Two
Fig. 8: Thingspeak Channel (Cloud Storage and Visualization) Fodder batches were tested with varying germination dates. It
was observed that best results for both batches were obtained
for different volume divisor values. For batch 1, volume divisor
V. E XPERIMENTAL S ETUP AND R ESULTS value of 4 gave 91.70% accuracy whereas for batch 2, volume
In this section the experimental setup and results are dis- divisor value of 5 gave 95.33% accuracy. It is assumed to be
cussed. Corn seeds were taken under consideration for the due to irregularity crop growth and uneven seed distribution.
prototype testing as it has high production efficiency, overnight Fig.8 shows Thingspeak log of values of various sensors along
seeds were allowed to sock in the warm water and then with estimated height, weight and price of the fodder batch.
taken out over a wet cloth and was covered to keep away Fig.9 shows, android application accessing and displaying
from sunlight for one more day.Further the seeds were taken values from Thingspeak Channel. Fig.9a shows all values in
out and spread evenly across the tray (1kg ),the trays were black font implying that all values are under desired range,
arranged as per the required cycle new trays were added into whereas Fig.9b, shows few values in red coloring implying it
the bottom shelf and each day the rotation was done it took be out of the desired range.
and days to grow in desired size once the cycle is completed As a side objective to determine overall operational cost,
constant socking and tray arrangement cycle were followed , the system was run for longer duration during which it was
 TABLE III: Operational Cost [2] C. J. G. Aliac and E. Maravillas, “Iot hydroponics management sys-
Hydroponics tem,” in 2018 IEEE 10th International Conference on Humanoid,
Traditional Nanotechnology, Information Technology,Communication and Control,
Input Output Requirement fodder
costing Environment and Management (HNICEM), 2018, pp. 1–5.
costing
[3] A. I. L. Maldonado, J. M. M. Reyes, H. F. Breceda, H. R.
84rs
1 kg 7 kg 3 trays/day 140rs/day Fuentes, J. A. V. Contreras, and U. L. Maldonado, “Automation and
/day
robotics used in hydroponic system,” Dec 2019. [Online]. Available:
140rs 56rs
1kg 10kg 3trays/day https://www.intechopen.com/chapters/70662
/day /day
[4] S. Salvi, S. A. Framed Jain, H. A. Sanjay, T. K. Harshita, M. Farhana,
N. Jain, and M. V. Suhas, “Cloud based data analysis and monitoring
of smart multi-level irrigation system using iot,” in 2017 International
observed that overall water consumption is less than traditional Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud)
method and the growth ratio is 7-8 times faster. Costing com- (I-SMAC), 2017, pp. 752–757.
parisons was studied and shown in Table.III,the setup output [5] M. V. Suhas, S. Tejas, Snigdha, S. Yaji, and S. Salvi, “Agrone: An
agricultural drone using internet of things, data analytics and cloud com-
study was considered for two average weight cow requiring puting features,” in 2018 4th International Conference for Convergence
1.5-2% of its body weight forage on daily basis ,the output in Technology (I2CT), 2018, pp. 1–5.
were categorized on the quality of corn seed germination one [6] R. Vidhya and K. Valarmathi, “Survey on automatic monitoring of
hydroponics farms using iot,” in 2018 3rd International Conference on
giving output of 7kg(Sr1) and other 10kg(Sr.2) ,The costing Communication and Electronics Systems (ICCES), 2018, pp. 125–128.
were taken from market analysis 1kg seed -28rs and traditional [7] H. K. Srinidhi, H. S. Shreenidhi, and G. S. Vishnu, “Smart hydroponics
mixture 8rs/kg+grains and other fodder costing 140rs per cow system integrating with iot and machine learning algorithm,” in 2020
International Conference on Recent Trends on Electronics, Information,
per day. Communication Technology (RTEICT), 2020, pp. 261–264.
[8] T. Kataoka, T. Kaneko, H. Okamoto, and S. Hata, “Crop growth esti-
VI. C ONCLUSION AND F UTURE W ORK mation system using machine vision,” in Proceedings 2003 IEEE/ASME
Low cost hydroponics system can be utilize using available International Conference on Advanced Intelligent Mechatronics (AIM
2003), vol. 2, 2003, pp. b1079–b1083 vol.2.
infrastructural and irrigation system which makes it cost [9] W. Bouachir, K. E. Ihou, H.-E. Gueziri, N. Bouguila, and N. Bélanger,
effective. A constant supply of green forage can be maintained “Computer vision system for automatic counting of planting microsites
improving the milk production capacity. Requires less man- using uav imagery,” IEEE Access, vol. 7, pp. 82 491–82 500, 2019.
[10] S. Boopathy, K. R. Gokul Anand, E. L. Dhivya Priya, A. Sharmila, and
power which inturn makes the system efficient for low income S. Pasupathy, “Iot based hydroponics based natural fertigation system for
farmers in underdeveloped countries as well as developing organic veggies cultivation,” in 2021 Third International Conference on
countries. Following are key points observed from results: Intelligent Communication Technologies and Virtual Mobile Networks
(ICICV), 2021, pp. 404–409.
• Low water usage: It takes 2-3 lit of water to produce 1
[11] S. Pawar, S. Tembe, and S. Khan, “Design of an affordable ph module
kg of green fodder where as conventional system takes for iot based ph level control in hydroponics applications,” in 2020
60-80lit and left over water can recycled. No manpower- International Conference on Convergence to Digital World - Quo Vadis
(ICCDW), 2020, pp. 1–4.
2-3hrs/day is the only requirement for such system to [12] D. Eridani, O. Wardhani, and E. D. Widianto, “Designing and imple-
work partially automatic. menting the arduino-based nutrition feeding automation system of a
• Reduce growth time: It takes 7-10days from seed ger- prototype scaled nutrient film technique (nft) hydroponics using total
dissolved solids (tds) sensor,” in 2017 4th International Conference
mination to fully grown plan of 25-30cm here conver- on Information Technology, Computer, and Electrical Engineering (IC-
sion ratio is 7-8 times that of traditional method. Any ITACEE), 2017, pp. 170–175.
sprouting plants can be consider majorly corn and wheat [13] M. F. Ali, P. Thakur, P. Mendiratta, and N. Gupta, “Iot-based solar hy-
droponics farming,” in 2019 6th International Conference on Computing
is taken. Constant supply of green fodder is maintained for Sustainable Global Development (INDIACom), 2019, pp. 927–931.
throughout the week. [14] D. Saraswathi, P. Manibharathy, R. Gokulnath, E. Sureshkumar, and
• Increase in nutrient value: It enhances milk production K. Karthikeyan, “Automation of hydroponics green house farming using
iot,” in 2018 IEEE International Conference on System, Computation,
quality as it helps synthesis Vitamin A. Completely Automation and Networking (ICSCA), 2018, pp. 1–4.
Natural and organic feed is being generated. [15] J. Li, Z. Mao, Z. Cao, K. Tei, and S. Honiden, “Self-adaptive hydropon-
• Weight and Price Estimation: Using image processing ics care system for human-hydroponics coexistence,” in 2021 IEEE 3rd
Global Conference on Life Sciences and Technologies (LifeTech), 2021,
techniques provided sufficient information to estimate the pp. 204–206.
price and weight of fodder. [16] C. Peuchpanngarm, P. Srinitiworawong, W. Samerjai, and T. Sunetnanta,
• Ease of Access: Ability to monitor vitals of Hydroponic “Diy sensor-based automatic control mobile application for hydropon-
ics,” in 2016 Fifth ICT International Student Project Conference (ICT-
System remotely using Mobile application provides im- ISPC), 2016, pp. 57–60.
portant information. [17] K. Tatas, A. Al-Zoubi, A. Antoniou, and D. Zolotareva, “iponics: Iot
In future the system can be improved by standardizing even monitoring and control for hydroponics,” in 2021 10th International
Conference on Modern Circuits and Systems Technologies (MOCAST),
distribution of seeds and providing same growth conditions. 2021, pp. 1–5.
More study on other fodder crops can be conducted with same [18] M. Talebkhah, A. Sali, M. Marjani, M. Gordan, S. J. Hashim, and
system to test its robustness and accuracy. Dynamic methods F. Z. Rokhani, “Edge computing: Architecture, applications and future
perspectives,” in 2020 IEEE 2nd International Conference on Artificial
needs to studied to estimate weight more accurately. Mobile Intelligence in Engineering and Technology (IICAIET), 2020, pp. 1–6.
application can be enhanced to control and manipulate factors [19] “Thingspeak for iot projects.” [Online]. Available:
affecting crop growth. https://thingspeak.com/
[20] 2021. [Online]. Available: https://appinventor.mit.edu/
R EFERENCES
[1] R. Gashgari, K. Alharbi, K. Mughrbil, A. Jan, and A. Glolam, “Com-
parison between growing plants in hydroponic system and soil based
system,” 08 2018.

View publication stats