Dissertation > Excellent graduate degree dissertation topics show

The Ship Detection Based on Optical Remote Sensing Images

Author: ShiPeng
Tutor: YuNengHai
School: University of Science and Technology of China
Course: Signal and Information Processing
Keywords: 3D reconstruction Saliency FIG. Gabor filtering
CLC: TP751
Type: Master's thesis
Year: 2010
Downloads: 243
Quote: 5
Read: Download Dissertation

Abstract

With more and more attention of the humans on the ocean, ships has become an important tool for people to take advantage of the ocean, the development of ocean. In order to better access to the information of the ship target ship target detection technology came into being, the applications are mainly focused on the various aspects of the relationship between national security, such as monitoring the waters, maritime search and rescue, fisheries, as well as counter-terrorism. With the traditional image detection algorithm based on synthetic aperture radar (SAR), because the characteristics of optical image imaging detection algorithm based on optical remote sensing images often ship target extraction more difficult, which is withdrawal of literature in recent years the problem by studying the imaging characteristics of the optical remote sensing image, we study focused on the following three aspects: an optical image because the light changes and showed brightness changes, making the area of ??marine and terrestrial two difficult segmentation algorithm through a single threshold to distinguish between . Optical remote sensing image there will be a lot of clouds interference factors, these different clouds shading, and similar parts of the ocean, the land, the cloud detection algorithm is more difficult to design. Optical remote sensing image appear complex sea conditions, combined with the ship target low reflectivity, making the the ship target difficult background distinguish. The emergence of the above problems will affect the ship target false alarm rate and missed alarm rate, based on extensive research literature on the above issues to focus on one by one, and designed for the optical remote sensing image ship target detection algorithm flow. The algorithm detection software in the form of the final application, GoogleEarth platform, compared to the ship target detection algorithm has a higher detection rate and low false alarm rate detection algorithm is proved by experiments. Research work and contribution of this paper is as follows: 1 for land and sea background brightness changes larger problem, propose an image-based 3D reconstruction of the land and sea segmentation algorithm, mainly the use of light and shade change information to reconstruct the image above sea level, combined with the morphological filtering and dynamic threshold segmentation algorithm to divide the land and marine areas. The experiments show that the algorithm is compared with traditional 2D threshold segmentation algorithm for land portray finer and higher robustness. 2 for chunks and imaging complex clouds detection problem, we use a method based on the image content feature learning feature vectors extracted from the gray scale, texture, edge angle describing the clouds, the re-use machine learning thought to be learning modeling using the model clouds classification. The last use of the idea of ??feedback algorithm, adaptively changing the parameters by adjusting the training set, by way of multiple classification makes clouds higher detection rate. Ship target detection in the context of complex sea conditions, we propose a detection algorithm based on human eye level mechanisms of visual perception and visual asymmetric. First, the use of an improved computing spectrum residuals visual saliency map; followed by the use of the the Tophat morphological filtering to remove visually does not belong to the area of ??the ship target and calculate interest concerns; Finally one direction adaptive Gabor filtering algorithms focus around a point of interest the association analysis, get one that has a clear ship target and the background noise is effectively suppressed image. In this paper, we summarize the full text, and discuss the outlook for the next step.

Related Dissertations

CLC: > Industrial Technology > Automation technology,computer technology > Remote sensing technology > Interpretation, identification and processing of remote sensing images > Image processing methods
© 2012 www.DissertationTopic.Net Mobile