Classification Model For Satellite Images Through ML

Satellite image processing is a highly developed field that still represents a great technical challenge. Machine learning methods are used for this purpose as there has been significant amount of research in the area of image processing through these techniques, and they offer the possibility of learning from a wide amount of data without supervision. In this project, we are developing a system to identify satellite photos so that information may be extracted using image processing techniques. The classification of satellite pictures into ten different classes, was completed. Used satellite photos are divided into residential, industrial, highways, Sea/lakes, rivers, permanent crop, pasture, herbaceous vegetation, forest, and annual crop. Manual classification utilizing picture interpretation techniques takes more time and requires the assistance of field professionals. As a result, we concentrated our efforts on effective automatic satellite picture classification. Satellite photos are classified, and features extracted using a convolutional neural network. CNN is a type of deep neural network that is ideal for dealing with images. CNN will assist in improving categorization accuracy. The overall categorization accuracy is estimated using a confusion matrix. We finally have achieved an accurate model that being trained/validated with almost 21,600 labeled images, reach an accuracy of 97.20% on the predicted output, being tested with a set of 5,400 images.