Design and Evaluation of a Deep Learning Model for Counting Objects in Satellite Images

Images obtained by satellites orbiting in space of the surface of the Earth or other celestial bodies are known as satellite images. These photos are an excellent resource for a variety of applications and have several uses. Applications are of earth observation, remote sensing and GIS etc. They are essential to tackling global issues and expanding scientific knowledge, and they play a crucial part in modern technology. Automated object counting in satellite images is a computer vision and image processing technique used to identify and quantify specific objects or features within images captured by Earth-observing satellites. After performing image preparation steps like using the CNN method and reducing noise, we must then recognize objects in the picture. Object counting has wide range of applications like Urban planning development, Military and Defence etc… In essence, automatic item counting in satellite photos helps businesses and researchers effectively obtain important information and make defensible choices. It considerably improves the capacity to manage and analyse extensive geographic data for a variety of applications.