A Discussion On Image Classification In Land Use Classification

Land use classification is a process of distinguishing land use types based on the way people use land resources,its purpose is to statistics and acknowledge the situation of land use so as to manage land resources,which is an important support for the decision-making mechanism of land management that ensure the accurate and rapid transmission of feedback in the decision-making system.Therefore,the issue of accuracy and efficiency has become an important factor that restrict the scientificity of the decision-making system.So it is of great significance to apply image classification to serve land resources management effectively.This paper studies the application of image classification technology in land use classification.Aiming at the common issue of accuracy and efficiency,the idea of constructing image classification model is proposed in order to improve the shortcomings in practical work.This paper puts forward the model integration method and use my own experience to designs two model architectures based on transfer learning which applied to land use classification: The first model is the Res Net-50 integration model and the second is ensemble model which combines VGG16 integration model and Xception integration model.This paper build a land database that refer to the glacier and water conservation ecological function area of Qilian Mountain,and use the models to carry on the empirical analysis of land use classification.The results show that theaccuracy of Res Net-50 integrated model is as high as 97.35% and the efficiency is 0.0318 seconds per sheet;the combined model has an accuracy of 96.57% and an efficiency of 0.026 seconds per sheet.Both of them have their merits,which can improve the accuracy and efficiency of land use classification and improve the decision-making level of land use management.