Research On Target Detection In Heavy Heterogeneous Clutter

Target detection under the background of the clutter is one of the basic applications in radar systems. With the development of modern stealthy technology, the radar cross-section(RCS) of the target significantly decreases, whose echo signal becomes extremely weak and the signal-to-clutter ratio(SCR) appreciably reduces. Meanwhile, the clutter generated by the illuminated scenarios including the urban area, mountain land, and sea waves, etc, behaves heavy and heterogeneous. These two factors jointly lead to the serious performance degradation of the traditional detectors designed under the assumption of the homogeneous clutter, especially in the case of the low SCR or heavy clutter, and also result in that the target detection against the heavy as well as inhomogeneous clutter is the austere challenge. Furthermore, it becomes the puzzle which must be solved in the signal and data processing of the modern radar.In this thesis, the problem of target detection under the background of the heavy heterogeneous clutter is addressed, and the relative research work includes clutter modeling, algorithm analyses, simulation assessment, and real-life data verification, etc, where the main content is shown as follows:1. In the heterogeneous environments of the heavy enclosed sea clutter and lake clutter, etc, modeled as the compound-Gaussian(CG) distribution with the inverse Gaussian(IG) texture, namely IG-CG distribution, the adaptive detectors are proposed in terms of the two-step generalized likelihood ratio test(GLRT) criterion. The detectors overcome the mismatch drawback of the clutter model in the existing non-homogeneous detectors, resulting in the improvement of the detection performance.2. For the inhomogeneous as well as heavy land/sea clutter with the partially correlated texture component, the two parts estimations of the texture component and the speckle component are employed to estimate the covariance matrix of the clutter, which adapts the detectors to the clutter according to the two-step GLRT criterion and reduces the influence of the mismatch of the clutter model on the detection performance.3. For the heavy urban clutter, sea clutter, and foliage clutter, etc, modeled as the heterogeneous CG distribution, the adaptive multiple-scan detectors(MSDs), namely GLRT-MSD, Rao-MSD, and Wald-MSD, are respectively proposed in terms of the two-step GLRT, Rao and Wald criterions. These MSDs effectively utilize the correlation diversity of the target and clutter across the multiple scans, hoped to improve the detection performance for the moving small target in the resolution cell.4. The multiple-scan non-homogeneous detector based on mean value(M-NHD) and that based on the steering vector(SV-NHD) by operating solely on the primary data are proposed under the heterogeneous background of cyclostationary sea clutter, which avoid the adverse effect on the detection performance of the moving target in the resolution cell caused by the heterogeneous secondary data.5. Considering the heavy urban clutter, sea clutter, and foliage clutter, etc, modeled as the heterogeneous CG distribution, the adaptive heterogeneous MSDs, VL-HSCD,VL-HKelly, and VL-HAMF, are proposed where the Viterbi-like(VL) cumulate algorithm and the hybrid estimation method of the covariance matrix are employed, which can improve the detection performance for the moving target in the disparate cells among the multiple scans.The mentioned detection approaches are tested on the simulation data or the measured real-life data, where the measured clutter data are from the commonly used IPIX radar in international and the clutter parameters of the simulation data are mainly from the estimations for the recorded live data.