Classification of remote sensing images using bp neural network with dynamic learning rate

Aim Backpropagation neural network classifier can solve the problems existing in the traditional classifers and has been gradually used in the classification of remote sensing image. A new improved BP method of classifying the remote sensing image is to be presented. Methods Conjugate gradient with line search (CGL) was introduced to optimize the learning rate. Results The training speed is much higher than other methods to save time from 5 to 110 s. Conclusion The method avoids the burden of the large storage and the divergence of the error function so that it is that it is applicable to remote sensing image classification.