Near real time object detection based on structured forecast edge method

Ship detection is an important issue in tasks related to sea traffic control, the observation of maritime activity has been a field of research for many years. It has been found that Satellite Synthetic Aperture Radar (SAR) is the only space-borne system that provides non-cooperative surveillance of maritime activities. In this thesis structured Forecast Edge method is used for finding a specific target over remote sensing images. A geometric object on a sea background from remote images is detected. These images can be influenced by clouds, waves, and other homogeneities. This regional method is obtained from edge detection, based on structured forecast, which makes our method accurate and efficient. This method helps in finding the ships in medium and high-resolution SAR images. Due to the highly variable environmental and weather conditions in closely spaced areas, achieving a controlled false alarm rate is a major problem. However, after various Edge Detection techniques and morphological processing on the contours, we have obtained a proposed shape by connected domain detections. Edge detection method is used for finding the sharp changes at the boundaries in an image, whereas the morphological process is used to remove the noise and stretches in the image for getting the exact and accurate result. Finally, SVM (support vector machine) method is used to determine the exact location of the ship