Research On Interference Suppression And Propagation Velocity Estimation In Ground Penetrating Radar

Ground Penetrating Radar (GPR) is an effective tool to detect the buried targets, and it's application is worldwide in many areas. The study of the GPR is one of the international researching hotspots and has become an important branch of the remote sensing technology. Because the GPR can detect the target rapidly, continuously and conveniently with high resolution, it has a broad prospect of application in many areas, such as national defense, anti-terrorism, civil engineering, construction of highway, railway, bridge, tunnel, prospecting, geologic studies, and archaeological studies.GPR is a very complex system, and its study is far from maturity as compared to that of traditional radar. Especially when GPR is used to detect the shallow buried target with high resolution, it must adopt the ultra wideband (UWB) technology and faces a series of new challenges to antenna design, signal modulation, signal sampling and processing. As the application of GPR goes wider and deeper, its working environment becomes worse and more complex, such as the increased radio frequency interference in the city, the more complex clutter and the influence of the half-space between the air and soil. All these require more consideration on both signal processing and system design. In order to optimize the whole performance of GPR, the signal processing is no longer a simple processing of the echo data of radar signal, it also requires the improvement of the radar system structure.The contents and innovations of this dissertation are as follows:1. A method of radio frequency interference (RFI) suppressing is proposed by using random equivalent time sampling in the impulse GPR. The property of the RFI after average filter is compared among real time sampling mode, periodic equivalent time sampling mode and random equivalent time sampling mode. The analysis indicates that the RFI can be transformed into zero-mean random signal by random equivalent time sampling without effect on the object's information and it can be suppressed easily by a classical average filter. Experimental results conform to the theoretic analysis and indicate the proposed method is effective.2. A RFI suppression method is proposed by using random phase codes in the stepped-frequency GPR. The random phase-coded method transforms the RFI into zero-mean random signal without effect on the object's information and makes the RFI be suppressed easily. It makes the stepped-frequency radar have the same Anti-RFI performance as that of the noise radar and reserve the low sidelobe performance. Monte Carlo simulations conform to the theoretic analysis and indicate the proposed method is effective.3. A novel clutter reduction method is proposed by using two-dimensional band-pass filter in frequency domain and symmetry filter based on the symmetry difference between the target reflection and the clutter in B-scan. The two-dimensional band-pass filter can effectively reduce the antenna coupling clutter, the ground clutter and the radar vibration noise. The symmetry filter can reduce the undulated ground clutter, soil roughness reflection and reflection signals from external anomalies. Experimental results confirm the effectiveness of the method.4. A fast method of propagation velocity estimation is proposed. The hyperbolic signatures and the apex in B-scan are extracted based on peak tracing and the symmetry of hyperbola. Moreover, according to the effect to velocity estimation in the apex section of the hyperbola, a weighted method of velocity estimation is proposed. The method is evaluated to be effective by using experimental results.5. A method of equivalent lateral velocity synthetic aperture imaging (SAI) is proposed. The object signal often is departure from the standard hyperbola in practice. The traditional algorithm using constant velocity can hardly amend the different. The proposed method amends it by using the equivalent lateral velocity which can be estimated quickly by using the hyperbolic signatures of echo in B-scan. The precise SAR imaging can be achieved by lateral variable velocity F-K migration. The experimental results demonstrate that the proposed method is more effective than constant velocity migration.