基于参数自适应CNN的倾斜摄影影像融合方法  

Oblique Photography Image Fusion Method based on Parameter Adaptive Convolutional Neural Network(CNN)

作　　者：刘明众[1] LIU Mingzhong(College of Geology and Construction Engineering,Anhui Technical College of Industry and Economy,Hefei 230051,China)

机构地区：[1]安徽工业经济职业技术学院地质与建筑工程学院,合肥230051

出　　处：《成都工业学院学报》2024年第6期39-44,共6页Journal of Chengdu Technological University

基　　金：安徽省教育厅高校优秀青年人才支持计划项目(gxyq2021279);安徽工业经济职业技术学院质量工程项目(2021xzytz01)。

摘　　要：针对目前遥感影像融合的精度较差的问题,搭建一种基于参数自适应卷积神经网络(CNN)的倾斜摄影影像融合方法。首先,搭建一种基于深度卷积神经网络的影像融合模型。然后,引入通道相似度注意力,以实现网络特征的自适应学习。结果表明,所提算法在空间相关系数方面的表现较好,指标为0.758,通用影像质量为0.57。此外,算法的影像失真较少,峰值信噪比为23.75 dB。在Pavia Center数据集中,所提算法展示出较好的空间细节信息保持能力、影像融合精度、匹配效果以及影像融合效果。在CAVE数据集上也具有较好的影像融合效果,相关系数最大,为0.997,光谱角反射和相对全局合成误差分别为4.440和2.617。实验结果证明,所提模型在影像融合方面具有较好的细节信息保持能力。To solve the issue of poor accuracy in remote sensing image fusion,a method of oblique photography image fusion based on parameter adaptive convolutional neural network(CNN)was studied and constructed.Firstly,an image fusion model based on deep convolutional neural network was built.Then,channel similarity attention was introduced to enable adaptive learning of network features.The results indicate that the proposed algorithm performs well in spatial correlation coefficient,with an index of 0.758 and a general image quality of 0.57.In addition,the algorithm has less image distortion and a peak signal-to-noise ratio of 23.75 dB.In the Pavia Center dataset,the proposed algorithm still shows good spatial detail information retention ability,image fusion accuracy,matching effect,and image fusion effect.It also has good image fusion performance on the CAVE dataset,with the highest correlation coefficient of 0.997,spectral angular reflectance and relative global synthesis error of 4.440 and 2.617,respectively.The experimental results demonstrate that the proposed model has good ability to preserve detailed information in image fusion.The research results can provide certain technical support for the image fusion of oblique photography and promote the development of oblique photography and remote sensing technology.

关 键 词：影像融合 卷积神经网络 自适应 倾斜摄影 注意力机制 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]