Object based Change detection on Temporal Multi-Spectral Satellite imagery

Monitoring changes in multi-spectral satellite images is important for tracking urbanization and environmental changes that happened on the earth. This study mainly focuses on usage of multi-spectral satellite images to identify the environmental changes such as natural world transformation. The proposed method in this study uses FCNN Algorithm and also LandTrendr for change detection with respect to Time Series data. Two distinct methods are used in this study to identify the changes in temporal and non-temporal data, respectively. The system's ability to identify changes in a variety of environments and landscapes is being assessed using evaluation metrics like accuracy, precision, and recall. The examined approach was evaluated over numerous satellite images and it was successful. This research makes it simpler to understand the changes in satellite images, which is very helpful for applications like city planning and environmental protection. This research offers a flexible framework that is useful for land use assessment, disaster management and urban planning. A useful method for monitoring dynamic environments, FCNN’s integration with object-based and pixel-based analysis improves the accuracy and usefulness of change detection. The LandTrendr algorithm gives complete changes with indexes like NDVI, NBI, etc. and it make use of Time series data (Temporal) in order to represent changed counts in the form a graph and measuring performance with Accuracy measure.