Research on Image Fusion Based on Interpolation Algorithm in Deep Convolutional Network Environment
Aiming at the shortcomings of the existing image fusion algorithms,an image inter-polation algorithm is proposed in the deep convolutional network environment.Firstly,the original im-age is preprocessed by filtering,segmentation,etc.,so that the image has the possibility of fusion;Then the deep convolutional neural network is constructed.In order to improve the training performance,the structure design of double convolutional layer and pool layer is adopted,and the Tanh activation func-tion is used to improve the generalization ability of the model.Based on variance gradient modeling and interpolation by neighborhood pixel method,the pixels of the fused image transition smoothly,strengthen the details of the fused image and improve the quality of the image.The experimental results show that the evaluation indexes of image fusion quality of the proposed algorithm are better than the traditional fusion algorithm,and the processing time of training set and test set is shorter.
deep convolutional networksinterpolationimage fusionfilteringneighborhood pixel method
万方数据
维普
