Ship Detection In SAR Images

With the development of the synthetic aperture radar (SAR) imaging techniques, the resolution of SAR images has been significantly improved. Using SAR images to detect ship targets has become the one of important applications in the field of ocean remote sensing.First, this paper introduces the reconstruction and the interpretation of SAR images, including the two-dimensional resolving of SAR images, the mechanism of speckle in SAR images, and the interpretation signs of SAR images. They are the theoretical basis of further study.Then, this paper analyses the factors influencing ship detection, discusses the basic knowledge about image segmentation and the basic methods of the ship detection in the SAR images, and tests and compares these methods by computer simulation.This paper presents an improved algorithm for ship detection in SAR images to further decrease the effect of speckle. Compared with existing algorithms, it has the following two improvements:1. Improved maximum entropy method. This method improves the original maximum entropy method according to the different distribution characteristics of ships and ocean surfaces in SAR images. The improved algorithm chooses the detection threshold to maximize the weighted sum of the background entropy and the target entropy. This method uses the weights to adjust the proportions of the background entropy and the target entropy in the criterion function, and yields a better detection threshold.2. Adaptive morphological processing. This algorithm finds the optimal size of the structuring element according to the probability density function of the speckle size, and thus has a better performance in smoothing the ships and suppressing speckle.SAR images with speckles with different intensities are simulated to evaluate the presented algorithm. The results show that this algorithm is effective even when the speckle is strong.Our country has a large marginal sea and rich marine resources. Ship detection in SAR images plays an important role in civilian and military applications. I hope that my work about ship detection in SAR images can contribute to our country's maritime surveillance and target detection technologies.