基于图像特征融合的遥感场景零样本分类算法

Image Feature Fusion Based Remote Sensing Scene Zero-Shot Classification Algorithm

利用不同图像特征之间的互补性,可提升遥感场景零样本分类性能。将图像特征的融合与零样本分类结合,提出一种基于图像特征融合的遥感场景零样本分类算法。采用解析字典学习方法,计算各图像特征的稀疏系数,并串接起来作为融合后图像特征,以减少冗余信息且保留各图像特征自身特点;引入监督信息,提高融合特征的鉴别性;将融合特征与场景类别词向量进行结构对齐,提升对新类别场景的迁移识别效果。在UC-Merced和航拍图像数据集两种遥感场景集上,对相同层次及不同层次的场景图像特征分别进行融合实验。实验结果表明:对于总体分类准确度和运算耗时,所提算法均优于其他零样本分类算法及通用的特征融合算法,证明了方法的有效性。

In order to improve the performance of remote sensing scene zero-shot classification by utilizing the complementarity among different image features, a zero-shot scene classification algorithm based on the fusion of image features is proposed, which combines the fusion of image features with zero-shot classification. The analysis dictionary learning is exploited to obtain the sparse coefficients of image features. The obtained sparse coefficients are concatenated as the fused image feature to reduce the redundant information and retain the characteristics of different image features. Supervised information is introduced to improve the discriminability of the model. The fused image feature is structurally aligned with the semantic word vectors to improve the transfer capability for the unseen class scenes. The fusions of image features at the same level and the different levels are carried out on UC-Merced(UCM) and Aerial Image Dataset(AID) remote sensing scenes datasets, respectively. The experimental results show that the overall classification accuracy and time-consuming of our method are superior to those of other zero-shot classification algorithms and general feature fusion algorithms, which proves the effectiveness of our method.