Font Size: a A A

Imaging Polarization And Terrain Classification

Posted on:2008-06-07Degree:MasterType:Thesis
Country:ChinaCandidate:D R WangFull Text:PDF
GTID:2208360212978912Subject:Pattern Recognition and Intelligent Systems
Abstract/Summary:PDF Full Text Request
It is an important subject in automatic target recognition to detect interesting objects from clutter background. Spectropolarimetric imaging has the ability to identify false targets and enhance images taken in poor visibility and even restore clear-day visibility of scene. In the thesis, spectropolarimetric imaging and materials classification based on spectropolarimetric imagery are investigated in detail. The main contributions are summarized as follows:1. In order to detect spectropolarimetric information exactly, an imaging spectropolarimetric system wad constructed. Liquid crystal tunable fi(?)er is utilized to change the wavelength of transmission light, and measure the polarized status in this system. This system is simple and st(?)ble. More than one hundred groups of spectropolarimetric images are collected by this system, and a small spectropolarimetric imagery database is set up.2. Some calibration approaches are given to make the imaging spectropolarimetricsystem's performance reliable, and to determine the relationship between the inputs and outputs of the system. 1b eliminate the disturbance caused by dark current and background light, a polarization retrieval method based on empirical line method is given. The experiment and simulation results show that the method is effective.3. To integrate the spectropolarimetric information in classification, an unsupervisedclassification fusion algorithm is proposed to improve classification accuracy The algorithm can be described as follow: first, three polarimetric parameter (intensity, degree of linear polarization and phase of polarization) imagery are clustered by fuzzy C-mean clustering(FCM). Then, basic belief assignment functions are constructed based on the cluster results and polarimetric characters of object. Finally, a fusion algorithm is implemented using weighted distribution meihod. The simulation results verify the effectiveness of the algorithm.4. The Spectropolarimetric Imagery Process System is designed. Polarimetric...
Keywords/Search Tags:Imaging Spectropolarimetric Detection, Stokes Vectors, System Calibration, Polarization retrieval, Unsupervised Classification
PDF Full Text Request
Related items