AI-Enabled Framework for Monitoring Salinity-Affected Soils in Coastal Andhra Pradesh using Remote Sensing Data and Deep Learning Algorithms
Implementing Organization
Siddhartha Academy Of Higher Education (Deemed To Be University)
Principal Investigator
Dr. ANURADHA GOVADA
Siddhartha Academy Of Higher Education (Deemed To Be University)
anuradhagovada@gmail.com
Project Overview
Soil formation takes more time and complex process to form one inch of top soil from parent material. Soil changes their properties based on climate, geomorphology etc. Remote sensing images play a vital part for the process of identification of different characteristics of soil that include texture, salinity, fertility and porosity and numerous more. Soil quality affects the crop yield the most. Soil texture, Soil Chemical Quality, Soil Physical Quality affect the soil quality. Among these several factors that affect the soil, salinity is the most important one which causes the major affect on soil quality. Therefore, identification of soil salinity helps to soil quality management methods. Soil salinization is, a natural as well as artificial phenomenon, defined as the huge accumulation of water soluble salts(NaCl) and at times some other compounds of potassium,sulfates etc. Soil salinization in majority of cases effects the growth of the plant which may lead to less vegetation or complete loss of farmland and may also lead to desertification (land degradation that reduces the fertility of land). The salinization of the soil inhibits the plant growth as plant healthy water diminishes even though soil has enough moisture. It will lead to death of the plant. It is important to identify the salt affected soils so that there will be no future agriculture practices on the saline land helping the farmers. The salinity of the soil can be identified using the spatial data (Remote sensed). Spatial data plays a vital role in the process of identification of different characteristics of soil that include texture,soil chemical quality,soil physical quality and many more. According to the studies, the soil salinity mapping on Indian soils using the Remote Sensing data has lot of scope. So, it is proposed to use remote sensing images for understanding the different kind of patterns that are not amenable normal human perception. The study area of this proposal is the state of Andhra Pradesh which mainly relies on the agriculture. The total 77.98 ha of irrigable land of Andhra Pradesh has been affected by salinization. So, we are identifying salt affected soils in a part of the Andhra Pradesh i.e. Krishna district and progress of salinization. The Krishna district, coastal area in the state of Andhra Pradeshis one of the best producer of irrigation based commodities.Remotely sensed data(landsat-8 ) coupled with Deep learning (Liquid Neural Networks) are proposed to be used for arriving at the the salinity indices from the collected images with the application of suitable mathematical technique.