Font Size: a A A

Application Of Empirical Mode Decomposition In Feature Extraction Of Remote Sensing Image Texture Direction

Posted on:2017-08-19Degree:MasterType:Thesis
Country:ChinaCandidate:X J LiuFull Text:PDF
GTID:2358330512468061Subject:Computer software and theory
Abstract/Summary:PDF Full Text Request
With the rapid development of spatial information technology, remote sensing technology and sensor imaging technology have made great progress.Remote sensing image information has gradually become our exploration of outer space, the observation of the change of surface features, it plays an increasingly important role in military reconnaissance, precision strike and civil facilities planning and construction.Remote sensing image feature extraction is the basis of remote sensing image processing? analysis and it is the guarantee to achieve a variety of observation functions, so the study of remote sensing image feature extraction has a very important practical significance. According to the traditional remote sensing image spectrum feature extraction methods in terms of accuracy affected by natural factors seriously affect the defects and remote sensing image itself characteristics, based on two-dimensional ensemble empirical mode decomposition algorithm and wavelet decomposition and Radon transform method to extract the features of remote sensing image texture direction, so as to make up for deficiencies of remote sensing image spectrum feature extraction.In addition, this thesis also for the two-dimensional set of empirical mode decomposition to produce small frequency than the letter composite modal mixed aliasing effect is not the problem, put forward understanding related to the two-dimensional ensemble empirical mode decomposition algorithm, and demonstrate its effectiveness.The main work of this thesis are as follows:(1) Analysis of the remote sensing image feature extraction research, summarizes the empirical mode decomposition (EMD) algorithm, the status quo at home and abroad; some methods and experience of the analysis and summary of the empirical mode decomposition algorithm to solve the end effect and modal aliasing phenomenon.(2) About the Mode Mxing Penomenon of EMD, EEMD is extended to two-dimensional, and the two dimensional ensemble empirical mode decomposition algorithm is obtained, and the feasibility and performance of the empirical mode decomposition algorithm are demonstrated by experiments.(3) For the extraction of remote sensing image texture features. Combined with the improved two-dimensional ensemble empirical mode decomposition algorithm and wavelet decomposition algorithm, to study the extraction and characteristics of remote sensing image texture direction, and the validity and superiority of the improved method is verified by experiment(4) Since the collection of empirical mode decomposition algorithm than dealing with small frequency mixed signal modal aliasing effect is not ideal; so we essentially produce modal aliasing of view, the introduction of understanding of the relevant empirical mode decomposition algorithm, and continue to expand to obtain a two-dimensional understanding of the relevant EMD reconciliation related two-dimensional set of empirical mode decomposition algorithm, and finally through the experiment proved the superiority after its improved algorithm.
Keywords/Search Tags:Empirical Mode Decomposition, Bi-Dimensional Ensemble Empirical Mode Decomposition, Texture orientation feature of remote sensing image, Bi-Dimensional Decorrelation Ensemble Empirical Mode Decomposition
PDF Full Text Request
Related items