eduzhai > Applied Sciences > Engineering >

Integrating global spatial features in CNN based Hyperspectral/SAR imagery classification

  • Save

... pages left unread,continue reading

Document pages: 4 pages

Abstract: The land cover classification has played an important role in remote sensingbecause it can intelligently identify things in one huge remote sensing imageto reduce the work of humans. However, a lot of classification methods aredesigned based on the pixel feature or limited spatial feature of the remotesensing image, which limits the classification accuracy and universality oftheir methods. This paper proposed a novel method to take into the informationof remote sensing image, i.e., geographic latitude-longitude information. Inaddition, a dual-branch convolutional neural network (CNN) classificationmethod is designed in combination with the global information to mine the pixelfeatures of the image. Then, the features of the two neural networks are fusedwith another fully neural network to realize the classification of remotesensing images. Finally, two remote sensing images are used to verify theeffectiveness of our method, including hyperspectral imaging (HSI) andpolarimetric synthetic aperture radar (PolSAR) imagery. The result of theproposed method is superior to the traditional single-channel convolutionalneural network.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×