Research On Enhanced Imaging Techniques Of Low-altitude UAV-Mounted UWB SAR

Ultra-wideband Synthetic Aperture Radar(UWB SAR) mounted on the low-attitude Unmanned Aerial Vehicles(UAV) has the characteristics of better soil or foliage penetration, higher resolution and richer aspect-frequency dependent information of targets, which make UWB SAR a promising method for detecting weak and small targets. However, detection of weak and small targets requires higher spatial capability and radiometric capability. Thus, the enhanced imaging techniques to enhance the spatial capability and radiometric capability of low-attitude UAV-mounted UWB SAR are investigated in this paper.Firstly, the real-time high resolution imaging techniques suitable for low attitude UAV-mounted UWB SAR are investigated. In order to get high range resolution, a new Radio Frequency Interference(RFI) suppression method is proposed based on the characteristics of pulsed stepped frequency wave. And then a data compression method based on 2 bit Block Adaptive Quantization(BAQ) is adopted on complex High Resolution Range Profile(HRRP) data. In order to get high azimuth resolution, the influences of system delay and time delays between the sensor data and UWB SAR echo are analyzed, and then an off-line delay-estimated approach based on the prepared targets is proposed. Further, the influence of lever error and the compensation method are discussed. Grid-BP is pointed out to be appropriate for paralleled acceleration and accelerated by multiple Graphics Processing Units(GPUs).Secondly, the influences of the error of geometry model on image capability are investigated. Aiming at the loss of image quality caused by attitude error, a compensation method combining the measured attitude data is proposed. To meet the requirements of the minefield detection, the BP imaging method intergrating with the Digital Elevation Model(DEM) data and geography coding is investigated. The high focusing qulity, small geometry distortion and accurate demarcator can be satisfied by imaging once and thus the processing flow is simplified. The concept of illumination is introduced to denote the illuminated energy per unit area. Then a novel approach integrating BP and radiometric correction is proposed. It has the advantage of restoring the radiometric distribution of the scene.Thirdly, aiming at the drawbacks of the traditional detection flow, the enhanced imaging theory and its framework is proposed and used in man-made targets enhancement and Body of Revolution(BOR) enhanced imaging. And thus the radiometric capability is improved. The aspect dependence and aspect correlation is analyzed by Finite Difference Time Domain(FDTD) simulated data. Then, a new method combining with the position and attitude error compensation is proposed to improve the precision of traditional subaperture technology. A manmade target enhanced imaging method based on aspect correlation is proposed to enhance the radiometric capability. The Aspect Scattering Entropy(ASE) is proposed to describe the diversity of aspect scattering and used in the BOR-enhanced imaging method. Then the application of the proposed method in landmine detection is discussed. The experimental results show that the enchanced imaging is effective to improve the radiometric capability of weak and small targets and thus the capability of detection improves.Finally, the aspect-frequency dependence and migratory scattering are integrated into the imaging methods which can enlarge the radiometric difference between targets of the same type but different sizes. A new enhanced imaging method using sparse representation is proposed based on the analysis of aspect-frequency dependence. The migratory scattering phenomenon is analyzed, and then landmine-enhanced visualization based on Circular Hough Transform(CHT) is proposed. The proposed method can also be used for other types of BORs.The primary research results achieved in this dissertation have been applied on the low-attitude Airship-Mounted UWB SAR(AMUSAR) system and thus prove that the obtained research results have well applicable value in practice.