Application of Deep Convolution Neural Network in Automatic Classification of Land Use

Abstract

Automatic classification of land use has always been a topic of concern for remote sensing and land science. It plays an important role in the field of land survey and land management and is the basis for the country to carry out land use planning. In last few years, with more and more high resolution remote sensing platforms is becoming usable, it is possible to update and evaluate land use classification quickly with the advantage of huge volume of data and more frequent of the image data updating. At the same time, we are facing more and more challenges of the big data in practice. With the rapid development and achievements of deep learning in the field of image recognition, this paper introduces a deep convolutional neural network to classify and evaluate the existing land use information, and conduct experiments and demonstrations through the self-constructed convolutional neural network. The test results show that the method has a good effect in the determination of houses, factories, greenhouses, waters and woodlands. Due to the small number of samples and the inconspicuous features, the site is confused with other land features, resulting in lower classification accuracy. The method of this paper can realize the automatic classification of land use types and the evaluation of classification effects.[1]