Remote Sensing Image Study Of Classification Based On Improved BP Neural Network

Remote sensing (RS) image classification is an important content of RS research area. How to solve the classification of the multi-category image and achieve certain accuracy requirement is a key question in RS application research, which has the very important practical significance especially in the geological survey and mineral exploration.The neural network technology is an important method of RS image classification. The BP neural network has a stronger learning capability and has a widespread application in the classifying. But, the BP neural network has shortcomings of the study convergence rate to be slow and the training process easy to fall into the partial minimum. Therefore, the paper improves BP neural network model by using the adaptive learning rate and the additional momentum, in order to enhance the classification accuracy of RS image.This paper designed improved BP neural network model based on the MATLAB neural network toolbox and researched the classification of 512×512 pixel experimental area of Mulei County of Xinjiang, by using the LANDSAT-7 ETM + RS image with the support of the project which investigated geology and mineral resources in the around 1:50 thousand regional Sepikou, Mulei County of Xinjiang in 2008.Through the error matrix for the precision analysis, it can be found that the overall classification accuracy which used improved BP neural network classification of RS image was 89.06%, Kappa coefficient was 85.53%. Compared with the maximum likelihood method, the traditional BP neural network method, the minimum distance method, the non-supervised classification method, the overall accuracy increased by 5.36%, 8.53%, 13.06%, 23.69%.In short, using improved BP neural network for RS image classification can satisfy the requirements of the regional geological work, process data quickly, and extract information, improving the ability to identify the wild. The method suits the western region or geographic areas of similar characteristics, similar to the units engaged in geological survey and related personnel, which can improve the work efficiency, save the funds and raise the informatization level.