Target segments of the Synthetic Aperture Radar (SAR) images are often of SAR images used as one of the challenge the procedure of SAR target recognition interpretation. Segmentation of the object, and then be separated from the target in order to eliminate the background noise or background clutter. However, segmentation may be discarded, and the shadow of the object where the target characteristic further comprises a portion of the difference between object recognition information. Thus, the proposed automatic detection and identification of target SAR image-based approach Sparse Representation Local Dictionary. (SRLD) Using the Model to target detection and recognition experiments to get moving and stationary targets, and identify baseline data set, verify the SAR image recognition field goal deep neural networks' effectiveness.

合成孔径雷达（SAR）图像的目标片段通常是SAR图像的一部分，是SAR目标识别解释过程中的挑战之一。分割对象，然后将其与目标分离，以消除背景噪声或背景杂波。然而，分割可以被丢弃，并且目标特征进一步包括对象识别信息之间的差异的一部分的对象的阴影。因此，提出了基于目标SAR图像的方法稀疏表示局部字典的自动检测和识别。（SRLD）使用该模型进行目标检测和识别实验，以获取移动目标和静止目标，并识别基线数据集，从而验证SAR图像识别领域目标深层神经网络的有效性。