Application of IMVR Convolutional Neural Networks to Classification of Land Use Remote Sensing Datasets

Yuanzhen Shuai, Ning Xin, Md Maruf Hasan, Bintao Hu, Tianhong Dai, Hengyan Liu

Abstract: Despite extensive research, remote sensing image classification remains a challenging issue within the field of remote sensing image analysis. Achieving a balance between classification accuracy and computational efficiency remains challenging, as traditional methods often face difficulties in attaining both high speed and precision simultaneously. To tackle this dilemma, we propose a method named IMVR which significantly reduces the computational burden while maintaining validity. This method enhances the richness and accuracy of high-dimensional feature representations through its output. Extensive experiments are conducted on the UC Merced Land-Use Dataset to demonstrate that our method can substantially improve classification performance and efficiency in comparison to traditional methods.