Font Size: a A A

Research On Key Technologe Of Ship Detection Based On Optical Remote-sensing Images

Posted on:2016-01-15Degree:MasterType:Thesis
Country:ChinaCandidate:S J LuanFull Text:PDF
GTID:2348330542976029Subject:Information and Communication Engineering
Abstract/Summary:PDF Full Text Request
In recent years,satellite remote sensing technology had a great development,large amount of high-resolution remote sensing images can be acquired.In civil and military fields,the use of optical remote sensing image automatically detecting ship targets has great application prospects.Research in this paper is automatic ship detection based on optical remote sensing images of complex scenes.The research is aimd to improve the adaptability and the detection efficiency of the system,focuses on the candidate ship target detection,ship wake detection and ship target discrimination.Firstly,Optical Remote Sensing Images were denoised,OTSU algorithm is used for land and sea segmenting.For candidate ship target detection,this paper introduces the two parameter CFAR algorithm.Considering the two parameter CFAR's shortcomes in optical remote sensing image ship detection,this paper introduces phase spectrum reconstruction based on DCT Ship for candidate ship target detection.By processing in frequency domain to detect small ship targets in sparse scene,this method makes up the shortcomings of dual parameters CFAR cannot detect dark ship targets.Compared to SAR images,ship wake feature in optical remote sensing images is not obvious,researchs on ship wake detection in optical remote sensing image is still in development stage,but the wakes in optical remote sensing image have richer grayscale and texture information,which can be used to target detection and discrimination.So this paper detects ship wakes after candidate ship detection,introduces improved CGHT line detection algorithm to detect wake feature quickly achieves better results in detection accuracy and efficiency.Finally,in ship target discrimination,this paper summarizes the existing studies,considers the diversity of false alarm targets,collects characteristics based on gray,texture and geometric of the candidate regions,introducts SVM classifier for target classification,and achieves ship target discrimination in complex scence.This paper uses a lot of seas optical remote sensing image with different characteristics for simulation,uses the results to verification,analyses and adjusts algorithms of ship detection system to ensure that the adaptability and effectiveness.
Keywords/Search Tags:Ship Detection, Optical Remote-Sensing Image Recognition, Visual Attention, Phase Spectrum, CGHT
PDF Full Text Request
Related items