Published Paper Details:

Synthetic aperture radar (SAR) is a notion that the creator of this project describes for detecting ships from satellite-taken sea photos. Using the Faster R-CNN (Region Based Convolution Neural Networks) Algorithm, ships may be located in SAR photos. The VGG ImageNet Network's ship photos will be used to train the RCNN algorithm, which will then extract features from the images based on the images' height, width, and colour channel. Convolution neural network (RCNN) filter train pictures features map from several layers. The train vector will be kept with every object detection from the picture with a value of 1, and all background features will be marked as 0. Every time a new SAR test image is uploaded, the RCNN algorithm will use the train vector to find objects with ships. The project's goal is to create a deep learning algorithm that uses SAR images as an input and processes it in segments. The application will show the same image in the console window when the procedure is finished, with marks if any objects that are ships are discovered; otherwise, the same image will be shown without any alterations. By conducting various experiments on our proposed system by collecting some sample ship images from KAGGLE website, we can train the application and then test the application on sample ship images. Here I used Google Collab as platform to execute the proposed application on some sample SAR images and find out the efficiency of our proposed application.