ABSTRACT

In the linear array synthetic aperture radar (LASAR) three-dimensional (3D) imaging, the spacing between adjacent elements in the uniform linear array (ULA) must satisfy the Nyquist sampling theorem to avoid the grating lobes, which makes the number of elements in the ULA very large. To reduce the elements in the ULA, the coprime adjacent array (CAA) with the same aperture length as the ULA is used when conducting LASAR 3D sparse imaging by compressed sensing (CS) algorithms. However, due to the increased autocorrelation coefficient of the measurement matrix, there exists grating lobes interference in the CAA-SAR imaging results. To solve this problem, we propose a joint sparse recovery (JSR) algorithm for CAA-SAR 3D sparse imaging. Firstly, we conduct sparse imaging on the CAA and its two subarrays, respectively. Secondly, the imaging results of the CAA and its two subarrays are performed image segmentation by the OTSU algorithm to extract their target-areas’ imaging results. Finally, we perform the image fusion by the wavelet transform on the target-areas’ imaging results to obtain the final imaging results. Both simulation and experimental results indicate that the imaging quality and computational efficiency of the JSR algorithm are higher than the random sampling array (RSA) and CAA under the same number of array elements. Besides, under the same aperture length, the JSR algorithm improves the computational efficiency than the ULA without imaging-quality loss.

摘要

线阵合成孔径雷达 (LASAR) 三维 (3D) 成像中，均匀线阵 (ULA) 中相邻单元之间的间距必须满足奈奎斯特采样定理才能避开栅瓣，这使得ULA 非常大。为了减少ULA中的元素，在通过压缩感知（CS）算法进行LASAR 3D稀疏成像时使用与ULA具有相同孔径长度的互质相邻阵列（CAA）。然而，由于测量矩阵的自相关系数增大，CAA-SAR成像结果存在栅瓣干扰。为了解决这个问题，我们提出了一种用于 CAA-SAR 3D 稀疏成像的联合稀疏恢复 (JSR) 算法。首先，我们分别对CAA及其两个子阵列进行稀疏成像。第二，CAA及其两个子阵列的成像结果通过OTSU算法进行图像分割，提取目标区域的成像结果。最后，我们通过小波变换对目标区域的成像结果进行图像融合，得到最终的成像结果。仿真和实验结果均表明，在阵元数相同的情况下，JSR算法的成像质量和计算效率均高于随机采样阵列（RSA）和CAA。此外，在相同的孔径长度下，JSR算法比ULA算法提高了计算效率，而没有成像质量损失。我们通过小波变换对目标区域的成像结果进行图像融合，以获得最终的成像结果。仿真和实验结果均表明，在阵元数相同的情况下，JSR算法的成像质量和计算效率均高于随机采样阵列（RSA）和CAA。此外，在相同的孔径长度下，JSR算法比ULA算法提高了计算效率，而没有成像质量损失。我们通过小波变换对目标区域的成像结果进行图像融合，以获得最终的成像结果。仿真和实验结果均表明，在阵元数相同的情况下，JSR算法的成像质量和计算效率均高于随机采样阵列（RSA）和CAA。此外，在相同的孔径长度下，JSR算法比ULA算法提高了计算效率，而没有成像质量损失。