Agriculture is the backbone of the country. A lot of factors such as climate change, population growth, food security concerns have driven the sector to seek more innovative/emerging technology/ approaches like AI and IoT's to improve crop yields and to get better farming results.
2. Use of Artificial Intelligence and Internet of
Things in Soil Science
Presented by,
Miss. Amruta D. Raut
Reg. no. Ph.D 2018/20
Seminar Incharge
Dr. B. D. Bhakare
Head
Department of Soil Science and
Agril. Chemistry
3. Outline of Presentation
1. Preamble
2. What is Artificial Intelligence
3. What is IoT
4. Importance of AI in IoT
5. Introduction to Drone or UAV
6. Application areas
7. Case Study
8. Conclusions
4. Being one of the oldest sectors and the backbone of the country,
developing the Agriculture industry has been a huge concern for the
Indian government.
A lot of factors such as Climate change,
Population growth, and
Food security concerns
Preamble
5. Artificial intelligence (AI) being a game-changer in other industries,
the Indian government has realised the importance and started to hold
this technology in developing the sector.
NITI Aayog - A National Strategy for Artificial Intelligence in India.
Aim - focusing on economic growth and social inclusion.
According to Agriculture Secretary Sanjay Aggarwal,
“AI and big data are going to be a game-changer in the agriculture
sector and the government is aiming to collate about 80% of such
data by 2020.”
6. • The term artificial intelligence was coined by the
American scientist John McCarthy in 1956.
• He defined it as the science and engineering of making
intelligent machines.
Artificial Intelligence
9. The Internet of Things (IoT) is the network of embedded devices,
which enables things to connect, collect and exchange data.
It is also referred to as Machine-to-Machine (M2M), Skynet
or Internet of Everything
10. System Architecture
Data from users
Cmd to actuator
Server
Show data to
users
Cmd from users
Mobile app
Web app
Send cmd to actuator
Sensor and
Actuators
11. Internet of Things
Huge amount of data
Artificial Intelligence algorithum
Data into useful actionable results
IoT devices
Importance of AI in IoT
3
4
5
2
1
12. Example - Soil Health Card
Big Data
SHC Card
Distribution of SHC 1.29 crore
(MH)
Soil Fertility Map
13. Maciej Kranz, Vice President of Corporate Strategic Innovation at
Cisco.
“Without AI-powered analytics, IoT devices and the data they produce
throughout the network would have limited value.
Artificial Intelligence and the Internet of Things is like a match made in Tech Heaven!!
14. Nikash App
• An Ingenious automated irrigation
system called Nikash which uses
IoT (Internet of Things) technology
to control irrigation in the fields.
Author - Duo, 2017
Controller App
Wireless
sensor
Soil
• Vijayeendra HS and Channabasappa Kolar’s Bengaluru-based
startup, Avanijal.
17. UAVs or Drones
The Flying Robots
No on-board pilot
Remotely controlled, semi-autonomous or autonomous
or combination
18. What is a UAS?
Unmanned Aircraft System (UAS)
- UAV
- Ground control station
- Pilot
- Visual observer
- Launcher
Unmanned Aircraft System (sUAS)- A system in which
the UAV weight less than 55 Ibs.
24. Limitation
Skilled and expert person
High Investment Cost
Large cultivated area
Government rules and regulation
25. Case study -1
• AI-sowing app by Microsoft (2016) and ICRISAT
• Study area - Devanakonda Mandal (Kurnool district - Andhra
Pradesh)
• Sample base -175 farmers
26. • In 2017, Same project expanded to 3000 farmers of
Karnataka and Andhra Pradesh during Kharif cycle
Crops include Groundnut, Ragi, Maize, Rice, and
Cotton, among others.
27. Case study -2
Soil Analysis and Monitoring
Study area - Raleigh, North Carolina, USA
(Sennaar et al., 2019)
Huge efficiency in the gains of agro-inputs by cutting the use of
chemical feritilizer nearly 40%.
The spatial analysis capabilities of GIS technologies helps in
efficient water management during irrigation.
28. Study area - Alfalfa (Riverdale, California)
(Fictchett et al., 2013)
GIS technologies in irrigation helped increase the per acre
crop output by up to 37.5%, and reduced water usage by
20%.
Source : AI in Agriculture
29. Case Study - 3
Crop Harvesting
• AI-enabled robots are being widely deployed on tomato
farms in Japan, and have reduced the on-field labour time by
20%
30. Case study - 4
IoT Enabled Soil Testing
(Department of computer science, mahe, MH, India and Tamilnadu)
Author - Sindhu P. and Indirani et al., (2018)
31. Fig. Block Diagram of Proposed System
NODE MCU
1.0
(ESP826612E)
wifi shield
Soil
Moisture
sensor
Humidity
sensor
Internet
Fig. Connection Diagram
Data IoT cloud server
(Thinkspeak.com) displayed data with
correct time and place.
32. Fig. Prototype of system
Arduino IDE is for wrting code.
Result
Data store in cloud cover. Graph
plotted in think speak as per
variation moisture, temperature and
Humidity level in soil.
Implementing this system will allow
users like farmers to monitor and
improve the productivity of the
crops.
33. Case study - 5
Monitoring of crop fields using multispectral and thermal
imagery from UAV
Author - Raeva et al., (2018)
34. Study area of Vysoke
Sensor : Multispectral and
Thermal sensor
Flight of
Drone
: Every month August to
April 2016
Drone : SenseFly fixed-wing
drone eBee
Software : Pix4D mapper
36. Fig. Barley: NDRE maps, May-July, 2016.
Index map
1. NDVI = NIR – RED/ NIR + RED
2. NDRE = NIR – RED edge / NIR + Red
edge
37. Case study -6
Use of vegetation index and meteorological parameters for the
prediction of crop yield in India
Author – Prasad et al., (2014)
38. Study Area :MaharashtraFig. Flow chart showing steps in building the crop yield prediction
model for wheat
Iterative and non linear quasi-newton
multivariate optimization for all variable
Loss function is least square i.e. Lf = (observed – Predicted)
39. Table. Prediction of crop yield for the years 1999,1998 and 1997
Equation
Maharashta (Rabi seaon)
Predicted(kg/ha) Observed(kg/ha)
b(1999) 1198.2 1288
b(1998) 1291.6 898
b(1997) 1164.1 1460
40. Case study-7
Application of Satellite Remote Sensing to find Soil Fertilization
by using Soil Colour
Author - Kumar et al., (2013)
41. Satellite Data - Landsat 8
Lattitute - 12° 20’ to 13° 20’
Longitude - 78° 10’ to 79° 40’
Location of Study Area (Vellore District)
Framework of soil identification method
using remote sensing
Landsat 8
Pre-processing
techniques
K- means Clustering
Image segmentation
Correlation
Soil Identification by colour
model
Processing
43. Case study- 8
Assessment of Soil Quality by Using Remote Sensing and GIS
Techniques
Author - Mohamed et al., (2018)
44. Fig. Location of study area
Study Area - Karnataka
Latitude - 11⁰ 40’58’’ and 12⁰
06’32’
Longitude - 76⁰ 24’14’’ and 77⁰
46’55’’.
Data - LISS –III and LISS –IV
SRTM
Data
Analysis
- ENVI software
Arc GIS
45. Physiographic unit of study area Land use/Land cover area of
Chamarajanagar district
Soil Quality Map
46. Case study - 9
Smart System Monitoring on Soil Using Internet of Things
(Vivekanandha College of Engineering for Women, Tamilnadu, India)
Author – Sowmiya et al., (2017)
47. Block Diagram of smart farm monitoring system
Microcontroller
P1C16F877 A
GSM
Buzzer
GPS
Soil Moisture
Temperature
pH rate of soil
Power Supply
48. Fig. Working Device in farm land
Sensor measure :
Temperature
pH range
Water level
Send the information to
registered users
49. Case study - 10
Performance evaluation of vegetation indices using remotely
sensed data
Author- Joshi and Chandra et al., (2001)
50. Study Area : Bhopal
Satellite Data : Landsat 5 TM
IRS P6 LISS IV
Average Elevation : 427 m
Average Rainfall : 1146 mm
Average
Temperature
: 25oC
Soil Type : Slightly deep,
well drained,
Calcareous
clayey soil
51. ERDAS Software Start ERDAS Imaging
Open the image in ERDAS and subset the study area
Georeferencing subset image using SOI Toposheet 55 E/8
Generate the Vis models for for both TM and LISS IV
data using modeler in ERDAS
Using Pseudo image of VIs image identify ranges of
land cover classes
Classify the data using Knowledge engineer classify
in ERDAS based on ranges for land cover classes
Range
==
Land cover
classes
Identify Pixels and assign class
Classified map with assign classes
Undefined pixels
Unsupervised class
Analysis
52. Results
Vegetation Index
Land cover class
Ladsat 5 TM IRS P6 LISS IV
NDVI SAVI NDVI SAVI
Water Body 1 1 1 1
Build-up land 0.722 0.6753 0.3506 0.4565
Open land 0.4483 0.5098 0.6753 0.7692
Sparse Vegetation 1 1 1 1
Dense Vegetation 1 0.8148 1 0.5833
Table. Comparison among responses of TM and LISS IV data for
different Vegetation Indices using K- value
53. Conclusions
1. AI in crop sowing has the potential to increase per acre crop
output as well as decrease input costs for farmers, analysing and
monitoring soil health helps to improve the sustainability of a
given piece of arable land and also it save the resources of
labour and time for farmers
2. IoT helps farmers to monitor and improve the productivity of
the crops.
3. Remote sensing technology helps in mapping, land use/land
cover changes, vegetation indices, also in predicting crop yield.
4. Drone play important role in precision farming like in mapping,
vegetation indices.