Soil Erosion Studies of Mangaluru Coastal Region Using Satellite Imageries and Machine Learning Algorithms

Coastal erosions usually cause drastic disasters in the ecosystems and the human lives in coastal zones. Change over time and identifying the location of the shoreline is the most important aspect of managing coastal areas and this requires frequent monitoring of the shoreline using satellite imageries over time. The effectiveness of filed-based study and image processing techniques for computing soil erosion or accretion rate are considered and the proposed method gives the precise results for soil erosion studies. In the proposed method, first principal component analysis (PCA) is applied, then image processing techniques are applied and finally, the model is developed using Machine Learning algorithms. Firstly PCA has been used for separating land and water bodies in satellite imageries, then image processing technique is used to compute soil erosion or accretion rate and finally, the model is developed using Machine Learning Algorithm. The proposed method is applied to satellite imageries, between 2014 to 2019, of the Mangaluru coastal region to attain erosion and accretion rate. The results showed that the erosion rate is comparatively high in Talapady, Someshwara, and Ullal Regions in the year 2014–2015. Also, for the years 2016–2018, there has been an erosion in these regions. Other regions of Mangaluru Coast are subjected to both erosion and accretion but there are no significant changes. As compared to field-based study and image processing technique-based soil erosion or accretion rate computation, the proposed method predicts the soil erosion or accretion for future years also. The model developed in the proposed study can process satellite imageries of several years and predicts soil erosion or accretion for future years. If the erosion rate is high then the government can take necessary action plans such as stopping sand mining, industrial and development activities and, planting more trees near the seashore.