Ccs National Institute of Agricultural Marketing: Automation Techniques in Irrigation

CCS NATIONAL INSTITUTE OF
AGRICULTURAL MARKETING
Automation techniques in irrigation

Presented by-
Group 6
AUTOMATION TECHNIQUES IN IRRIGATION
AND ENABLING FARMERS
• India follows traditional agricultural methods in irrigation practices. Irrigation is a significant
factor in determining the crop yield and largely varies with the geographical, climatic, and
topological factors. Farmers primarily depend on personal monitoring and their experience in
irrigating the fields, and as a result, irrigation becomes largely inefficient and irregular. India,
therefore requires a simple irrigation solution on which the farmers can depend indefinitely, which
can adapt to the local climatic conditions, and accurately predict the quantity of water required by
the crops in real time to ensure judicious use of water resources, and also a better crop yield.
• Irrigation is one of the most labor intensive processes in farming. AI trained machines aware of
historical weather pattern, soil quality and kind of crops to be grown, can automate irrigation and
increase overall yield. Nearly 70% of the world’s fresh water resource is utilized for irrigation;
such automation can conserve water and benefit farmers in managing their water probs.
HOW IT WORKS
1. we make use of Node MCUs, and soil moisture sensors, placed inside waterproof boxes and
spread evenly throughout the area to be irrigated. All these nodes are connected to a Raspberry Pi
3B+ via wireless LAN.
2. The system analyses soil moisture content through the deployed sensors, which is used to predict
the quantity of water required to irrigate the area with an appreciable accuracy. It is a highly
autonomous system which requires little to no human intervention once deployed in a field.
3. The system developed makes use of Random Forest Regressor to predict the weather. It makes
use of the traditional data of rainfall patterns and weather data.
4. It has the ability to slowly adapt to the region-specific climatic conditions and its accuracy
improves with every prediction made. The system is designed to be updated every 30 minutes as a
small regular interval which makes it power efficient as well.
METHODOLOGY
The system designed consists briefly of two major components:
1. Machine Learning to predict the amount of rainfall and the soil moisture content
2. Economic Hardware Implementation using Internet of Things.
• Central Controller
• Distributed Nodes
• The Network
• Rainfall Prediction using Machine Learning
• Estimation of Soil Moisture Content required
1. Central Controller
A central system is used to control the nodes, and to decide the time for which the pump
attached to any particular node is switched on. This implementation uses a Raspberry Pi 3B+,
running the Debian Linux-based Raspbian Operating system as the central controller. The
functions include communication with the nodes, data preprocessing and deriving inferences
using Machine Learning algorithms.
2. Distributed Nodes
Each node is comprised of an ESP8266 based microcontroller board (NodeMCU) and an
array of analog sensors. These sensors are used to fetch information about weather
conditions, which is converted to digital 10-bit integers using the Analog to Digital converters
present on each NodeMCU. An algorithm converts this data to a 32-bit floating point value
required by the inference model, on the NodeMCU itself
3. The Network
The nodes are connected to the central controller through a shared Wi-Fi access point,
over the IPv4 protocol . A client-server architecture is used for communication between
the nodes and the central controller. The nodes do not communicate with each other. This
implementation uses a Flask (Python) server on the central controller. The server extends
a URL in the form of a webhook. The nodes connect to the server via this webhook to
send data. The HTTP GET request is used to transmit the sensor data after post-
calculation to the server.
The default response of the server conveys two messages to the node that sent the
request:
1. If the pump attached to the particular node has to be switched on or not, and
2. The duration for which the pump has to be switched on.
4. Rainfall Prediction using Machine Learning
The rainfall prediction involved a two-phase solution:
● Prediction of Probability of Rainfall in the next 30 minutes
● Estimation of the Amount of Rainfall
The first phase is to help the network to realize whether there is a chance of rainfall to occur in the next 30
minutes or not and the device keeps checking the status at regular intervals of time when it is switched on
periodically. This is achieved by predicting the probability of rainfall to occur in the next 30 minutes.
5. Estimation of Soil Moisture Content required

The decision system takes into account the water level content required by the crop to calculate its irrigation
requirements. In traditional applications, these decisions are taken by the farmer based on their experience or an
expert agricultural expert. Information from various sources is collected; these include weather statistics, crop
and soil properties, moisture content data collected from soil sensors deployed in the fields, and moisture
influencing factors such as evapotranspiration. An agricultural expert analyzes this data to decide an estimate of
the amount of water that the crop would require on a particular day.
CONCLUSION
• The system designed is a smart irrigation solution based on artificial intelligence
which is makes use of the soil moisture content and the moisture requirements of
the crop to make the entire process of irrigation automatic. Its core benefit is its
efficiency and economic feasibility. The main idea behind the rainfall estimation
proposed is to get the almost correct estimation of the rainfall of a specific local
region as well as the annual rainfall to get data for future estimation of the rainfall
in states.
• The autonomous irrigation system we developed uses machine learning and

predictive algorithms to add intelligence to existing concept of automatic
irrigation systems. The methods discussed in this paper can help increase
irrigation efficiency while decreasing effort required and aids water conservation

Ccs National Institute of Agricultural Marketing: Automation Techniques in Irrigation

Uploaded by

Copyright:

Available Formats

Ccs National Institute of Agricultural Marketing: Automation Techniques in Irrigation

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ccs National Institute of Agricultural Marketing: Automation Techniques in Irrigation

Uploaded by

Copyright:

Available Formats

CCS NATIONAL INSTITUTE OF

Automation techniques in irrigation

5. Estimation of Soil Moisture Content required

• The autonomous irrigation system we developed uses machine learning and

You might also like