Analysis of Agriculture Production and Impacts of Climate Change in South Asian Region: A Concern Related with Healthcare 4.0 Using ML and Sensors

Abstract

The Effect of Global Warming and rapid changing climate in an indefinite manner is a major concern and all domain of science are trying to address it in their ways. It is not only creating challenges to food production, yet to the human health. The presented research work is all about the prediction of the yield of agriculture of the land without involving any activity of humans and this makes our procedure superfast and quite easy and reliable for humans and hence the name of the project “Predicting Agricultural Productivity”. Main purpose of the research work includes the implementation and training of machine learning algorithms for the prediction of the yield of agriculture so that the error can get minimized and accuracy gets maximized. For training of the model, a collection of features from actual yield and pictures of satellite is extracted by us. After This phase, a suitable algorithm like Naive Bayes, NN and its variant are chosen and used as the mathematical way to learn the parameters that are based on the features of yield. Then during study, harvest of agriculture is prognoses for a separate set of data. Data that is prognosticated is compared in contrast to the actual land yield. The manuscript also focuses the different data sets which are obtained by satellite imaging and using remote sensing, the clear mapping of current condition is obtained which helps to predict the yield in better way.

Appendix

Data Format

Following is data used for testing and training has particular features listed below:

• Id

• State: It basically is used in order to find out the place of which the user wants to predict the agricultural yield.

• Crop: It is basically used to determine whether it is wheat or rice or maize etc.

• District: It basically is used in order to find out the place of which the user wants to predict the agricultural yield.

• Year: It is used to know that which year’s data need to be accessed.

• Season: It is asked to user to know that which season’s data need to be accessed.

• Area in hectare

• Production in tonnes

• Yield in tonnes/hectare

• MODIS Image Array.

TRMM Image Array.

Input Features: Satellite Data

a-MODIS: The Moderate Resolution Imaging Spectro radiometer is a payload imaging sensor built by Santa Barbara Remote Sensing that was launched into Earth orbit by NASA in 1999 on board the Terra Satellite, and in 2002 on board the Aqua satellite.

b-TRMM: The Tropical Rainfall Measuring Mission was a joint space mission between NASA and the Japan Aerospace Exploration Agency designed to monitor and study tropical rainfall. The term refers to both the mission itself and the satellite that the mission used to collect data.

• A. MODIS satellite data are available twice daily at 500 m resolution for any area in the world from 2000—present.

• B. TRMM satellite data are available every 3 h at 25 km resolution for any area in the world from 1997—present (Figs. 3.11, 3.12, 3.13, 3.14, 3.15, 3.16, 3.17, 3.18 and 3.19).

Fig. 3.11

Sample APY data

Fig. 3.12

Sample NDVI and TRMM images (Anantpur)

Fig. 3.13

India LULC data

Fig. 3.14

Design and methodology at a glance

Fig. 3.15

Source National Remote Sensing Centre

Domains in agriculture.

Fig. 3.16

Global vegetation greenness

Global Vegetation Greenness: Vegetation greens can be determined by using remote sensing. Red and grey are regions with low NDVI below 0.1 and represent barren sand, rock or snow. Yellow and light green are regions with moderate NDVI between 0.2 and 0.3 and represent shrub and grassland. Dark green are regions with high NDVI and represent tropical rainforests.

Fig. 3.17

Sun reflectance model

Determine the amount of reflectivity.

• 1. The sun’s energy is absorbed, reflected or re-emitted by vegetation depending on the amount of chlorophyll in the plant.

• 2. Satellites collect the reflected or emitted energy from the vegetation.

Fig. 3.18

Vegetation reflectance model

In green vegetation:

• Chlorophyll absorbs most of the visible ed and blue to make food

• Much of the near infrared is reflected.

Fig. 3.19

Normalized difference vegetation index (NDVI)

Introducing a Normalized term:

• Normalized Difference Vegetative Index is a remote sensing method that allows one to display greens of vegetation.

• NDVI compares reflectivity of NIR and Red wavelength bands.

• Image analysis using remote sensing allows one to determine NDVI using this formula (Fig. 3.20):

  $$ {\text{NDVI}} = \left( {{\text{NIR}} - {\text{Red}}} \right)/({\text{NIR}} + {\text{Red}}) $$

Fig. 3.20

Crop yield prediction working