Exploring data analytics techniques for the conservation of natural resources using spatial data

Abstract

Land use land cover (LULC) usually alludes to the assortment and cataloging of certain activities carried out by humans together with the natural elements on the land. Sentinel satellite images are meant to obtain optical images at high spatial resolution say of about 10m. Here, we have used three predominant bands namely NIR, Red and Green to classify the sentinel data with five classes namely Water, Forest, Vegetation, Urban and Open land of Bangalore region, Karnataka, India. Also, classified maps are generated using different neural networks with pixel-based classification approach. For, the proposed dataset, an inclusive accuracy of 95% was achieved with deep neural networks compared to various deep convolutional neural network architectures such as ResNet152V2, MobileNetV2, EfficientNetB0. newline Further, we aimed at producing optimal land use land cover map with various band combinations of sentinel satellite imagery as they represent different characteristics of spatial data obtained from google earth engine. Convolutional Neural Network technique outperformed with 98.1 % of accuracy and less error rates in confusion matrix considering RGBNIR (4328) band combination of satellite imagery. newlineThe chosen study area is more prone to urbanisation and greatly affected by population in recent years. Spatial-temporal data from 1989-2019 are considered. An optimal LULC maps from 1989 to 2019 obtained by deep neural network technique are used to perform change analysis which would mainly give the change LULC map with number and percentage of change pixels. According to the analysis performed major change as environmental affecting factor was noticed between 2009 and 2019 where in urban with the area of 189.3861 sq. km remain unchanged and noticeable transitions from other LULC classes to urban. newlineTime series classification was performed using Cellular Automata, Cellular Automata-Neural Networks, techniques to predict the LULC map of 2024.