Font Size: a A A

Research On High Precision Registration Method Of Millimeter Wave Radar Tomographic Image

Posted on:2022-12-09Degree:MasterType:Thesis
Country:ChinaCandidate:Y T LiFull Text:PDF
GTID:2518306764979389Subject:Automation Technology
Abstract/Summary:PDF Full Text Request
Millimeter-wave radar has the advantages of large bandwidth and small antenna aperture,and can finely distinguish target features.The millimeter-wave radar tomography system obtains multi-track sub-image sequence sets in the height direction.Uses the phase difference of the target at different heights to synthesize the virtual aperture to obtain the height resolution capability,and realizes the high-precision 3D imaging of high-value targets,which is important in the fields of security inspection and so on application value.Due to the inevitable differences between trajectories and platform motion errors in the millimeter-wave radar tomography system,the two-dimensional images of different heights have positional differences.Whether the multi-track sub-image sequence set is registered and the phase is preserved is one of the decisive factors for the height and resolution of the 3D imaging results.Taking the near-field orbit millimeter-wave radar as an example,this thesis conducts theoretical analysis and experimental verification around the echo model of the millimeterwave radar tomography system and the registration of tomographic image sequence sets.The main research contents are as follows:1.Construct the geometric and echo signal model of the millimeter-wave radar tomography system,analyze the trajectory error between the tomographic sub-images,and lay a theoretical foundation for the subsequent research on the millimeter-wave radar tomographic image registration method.2.Study the realization mechanism of the registration method,establish the image transformation model,and analyze the affine transformation model suitable for global image registration.3.A phase-preserving registration method based on sub-image trajectory error estimation is proposed.The proportional gradient calculation method is used to improve the accuracy of image feature point extraction and feature point description.The feature points in multiple sub-images are matched to obtain registration.According to the correlation analysis between the sub-image trajectory error and the registration offset,the linear error trajectory is corrected.Finally,the high-precision registration of the millimeter-wave radar tomography image is realized and the phase information is preserved.4.A phase-preserving registration method based on attribute scattering center estimation is proposed.The fast self-iterative shrinkage threshold algorithm is used to estimate the model parameters of the decoupled attribute scattering center.Based on the model parameters,a point feature descriptor is established,and the sub-images are registered.After obtaining the registration offset,the linear error trajectory is corrected.Finally,the highprecision phase-preserving registration of the millimeter-wave radar tomography system is realized.The above models and methods are verified by simulation data and measured data.The experimental results show that the high-precision registration method of millimeterwave radar tomographic images proposed in this paper can effectively realize the tomographic envelope alignment of tomographic image sequence sets.High-precision coordinate alignment between domains,and phase information is preserved.
Keywords/Search Tags:Millimeter-Wave Radar, Sparse Tomographic System, SAR Image Registration, Systematic Error Estimation, Attributed Scatter Center
PDF Full Text Request
Related items