Paper/Poster/Presentation Title

Artificial Intelligence for Intelligent Agriculture

Presenter/Author Information

Ramana Lingampally

Keywords

Participatory modelling, deep learning, seed germination, artificial neural networks

Start Date

16-9-2020 4:20 PM

End Date

16-9-2020 4:40 PM

Abstract

Low crop yield is a growing concern all over the world. One of the reasons for this is lack of precise knowledge and preparedness among farmers about different factors that affect crop yield. Worsening weather conditions and ecological imbalances only made the matters worse. In a country like India where small and marginal farmers account for more than 80% of the farming community face several challenges with rising temperature and unpredictable rainfall. Many a time seed germination process takes a hit resulting in total crop failure or low crop yield. Numerous studies have been carried out to improve the crop yield, seed quality and seed germination. After reviewing more than 40 research papers, it is clear that different machine learning models developed primarily focused on nutrients and seed quality but not much involvement of farmers and environmental factors. Mostly regression technique is used for precision agriculture and K-NN for seed quality. Artificial Neural Networks, Deep Learning and Support Vector Machines are used for classification and regression to identify diseases and growth. But most of these algorithms won’t work well in the fields. To illustrate this more, water required for a crop depends on temperature, humidity, soil type and type of crop. Taking this reality into account and after conducting some studies and taking farmers’ input, it has become clear that multiple interactions with farmers would be of immense help in seed germination process a success. Participatory modeling (PM) for farms helped include many challenges including environmental factors in the data. Data, qualitative and quantitative, from 20 different farmers at the seed stage was prepared and analyzed using data science and data visualization techniques. Temperature, moisture, air, and light conditions must be correct for seeds to germinate. The visualizations made it easy for farmers to understand different patterns during germination of seeds. The feedback given by them helped come up with new models. As the data available was limited, transfer learning and deep learning were used to analyze come up with different seedling growth patterns for different types of seeds. This definitely is of help to farmers as many a time, crop failure occurs at the germination stage itself.

Stream and Session

false

Share

COinS
 
Sep 16th, 4:20 PM Sep 16th, 4:40 PM

Artificial Intelligence for Intelligent Agriculture

Low crop yield is a growing concern all over the world. One of the reasons for this is lack of precise knowledge and preparedness among farmers about different factors that affect crop yield. Worsening weather conditions and ecological imbalances only made the matters worse. In a country like India where small and marginal farmers account for more than 80% of the farming community face several challenges with rising temperature and unpredictable rainfall. Many a time seed germination process takes a hit resulting in total crop failure or low crop yield. Numerous studies have been carried out to improve the crop yield, seed quality and seed germination. After reviewing more than 40 research papers, it is clear that different machine learning models developed primarily focused on nutrients and seed quality but not much involvement of farmers and environmental factors. Mostly regression technique is used for precision agriculture and K-NN for seed quality. Artificial Neural Networks, Deep Learning and Support Vector Machines are used for classification and regression to identify diseases and growth. But most of these algorithms won’t work well in the fields. To illustrate this more, water required for a crop depends on temperature, humidity, soil type and type of crop. Taking this reality into account and after conducting some studies and taking farmers’ input, it has become clear that multiple interactions with farmers would be of immense help in seed germination process a success. Participatory modeling (PM) for farms helped include many challenges including environmental factors in the data. Data, qualitative and quantitative, from 20 different farmers at the seed stage was prepared and analyzed using data science and data visualization techniques. Temperature, moisture, air, and light conditions must be correct for seeds to germinate. The visualizations made it easy for farmers to understand different patterns during germination of seeds. The feedback given by them helped come up with new models. As the data available was limited, transfer learning and deep learning were used to analyze come up with different seedling growth patterns for different types of seeds. This definitely is of help to farmers as many a time, crop failure occurs at the germination stage itself.