姿态图像缺失情况下的SAR目标识别

针对目标姿态图像缺失的情况,提出通过姿态图像合成的方式增加训练集的姿态覆盖程度,并将扩充后的图像也用于训练目标分类器.受稀疏表示模型的启发,建立了一种合成孔径雷达图像姿态合成模型.该模型根据少量已知姿态的图像,线性组合出缺失姿态下的近似图像.在运动和静止目标获取与识别数据集上的实验表明,通过合成缺失姿态下图像的方法可有效提升目标识别的精度,特别是在训练数据集中姿态缺失严重时,文中方法提升尤为明显.

SAR image target recognition in lack of pose images

The performance of synthetic aperture radar (SAR) image target recognition depends on the diversity of pose images in the training set. The problem of lack of pose images is considered, and the method of training data augmented with the synthesized pose images is introduced to train the classifier for target identification. Inspired by the sparse representation model, the model for synthesizing pose images is also developed, which approximately construct the missing pose image by linearly combining several images available. Experimental results on the moving and stationary target acquisition and recognition (MSTAR) dataset show that the proposed method of pose images synthesis can increase the recognition accuracy effectively. In particular, significant improvement can be obtained in the case of serious lack of pose images.