Development of Land Cover Classification System Implementation Utilizing an Artificial Neural Network (ANN) Classifier

"Land usage" defines how that region is used, whereas "land cover" relates to the planet's surface. Among the various forms of land cover are water, snow, grasslands, deciduous forests, and bare soil. This research proposes utilizing an artificial neural network (ANN) classifier to implement a land cover categorization system. First, PCA (Principle Component Analysis) is used to preprocess the input image in order to reduce its dimensionality. The picture that has been preprocessed is then sent to feature extraction. For feature extraction, a convolutional filter is employed. It is possible to extract statistical characteristics like minimum, maximum, and standard using this feature extraction procedure. These features are trained and tested using the extracted features. Go on to the image of the ground truth next. This ground truth picture includes example photos of things like buildings, lakes, and agricultural regions. The next step involves creating a classification model by dividing these images into training and testing sets. Land cover regions are accurately predicted as a consequence of the retrieved image's effective classification by an ANN classifier. MATLAB simulation software is used to carry out this research. The Accuracy Comparison of the ANN is 91%, Specificity Comparison of ANN is 91.5% and Sensitivity Comparison of ANN is 93.5% respectively.