ID 原文 译文
5734 该算法引入多个候选集,通过设定阈值来进行原子筛选和候选集数量的调整。 The algorithm is introduced into multiple candidate sets, by setting the threshold for atom selection and adjustment of the number of candidates.
5735 然后每次迭代选出残差最小的路径作为新的候选集,以提高重构速度。 Then choose the path of least residual at every iteration as a new candidate set, in order to improve the reconstruction speed.
5736 此外,将残差差分小于某一阈值作为算法停止条件,因此不需要图像稀疏度作为算法的输入。 In addition, the residual difference is less than a certain threshold as a condition of algorithm to stop, so there is no need to image sparse degree as the input of the algorithm.
5737 实验结果表明,该算法可以获得较好的重构效果,同时保持了良好的时间复杂度和抗噪性能。 The experimental results show that the algorithm can obtain better reconstruction effect, while maintaining the good time complexity and antinoise performance.
5738 对地目标识别中,目标回波中将不可避免的含有大量杂波,严重影响识别性能。 To ground target recognition, target echo will inevitably contains a lot of clutter, seriously affect the recognition performance.
5739 针对该问题,提出了杂波稳健识别方法,该方法首先采用基于球形假设聚类的异常检测方法,完成小簇杂波的抑制; Aiming at the problem, put forward the clutter robust identification methods, this method first USES the anomaly detection method based on spherical assumption clustering, complete the small cluster of clutter suppression;
5740 然后利用参数非依赖局部异常因子异常检测法对剩余散射点进行异常检测,完成点杂波的抑制; Then using parameters abnormal test not rely on the local anomalies factor to anomaly detection of remaining scattering points, completion point clutter suppression;
5741 最后利用杂波抑制后的高分辨距离像进行特征提取和目标识别。 After the last use of clutter suppression of high resolution range as the feature extraction and target recognition.
5742 本文方法对目标是否运动无要求,算法涉及参数不需要人为设定。 Whether the method of target movement no requirement, the algorithm involves parameters do not need artificial.
5743 实测数据验证表明,所提方法的识别率优于传统方法,随着信杂噪比降低识别性能稳健,且运算量较小。 Measured data validation show that the proposed method recognition rate is superior to the traditional method, with the letter of mixed noise ratio reduce the recognition performance is steady, and the computation is small.