面向桥梁目标自动检测的多源遥感图像融合模型与方法

Fusion Model for Bridge Automatic Detection in Multi-source Remote Sensing Imagery

作为信息增值的一种有效手段,图像融合技术已广泛地用于成像遥感应用领域。面向目标识别的图像融合需依据不同的目标任务建立不同的融合模型。该文提出了适于桥梁目标自动检测的多源遥感图像融合模型和方法。该模型充分考虑了数据源特性、遥感图像中的桥梁目标特征和目标背景特征,能在复杂大幅场景中有效地检测桥梁。真实多源遥感图像的实验结果验证了该模型和方法的合理性。

Image fusion principles have been widely used in application of imaging remote sensing as an effective means of synergistic combination of information. Image fusion applied for target recognition should be modeling for each different task. A kind of fusion model for bridge automatic detection in multi-source remote sensing image is proposed in the paper, which is designed based on the character of multi-source data, bridge identity and the background feature in remote sensing images. The results of bridge automatic detection in real images of complex scene prove the model＇s validity.