A Sub-Pixel Mapping Method Based On Logistic Regression And Pixel-Swapping Model

Mixed pixels are widely existed in remote sensing data. Using the proportion of different land-covers to improve the spatial resolution of hyperspectral images is a popular method in the field of remote sensing data processing. The proportion data and location of sub-pixels in geometrical shapes can be used as the training data to train the neural network. The trained model can be used to sub-pixel mapping for the real land image. This paper proposed a sub-pixel mapping method based on Logistic Regression and Pixel-Swapping Model (LRPSM). The artificial image and real land image taken by Landsat8 were used to be tested. Experiments showed that the accuracy of LRPSM outperformed PSM based on sub-pixel spatial attraction model and BPNN based on neural network model in sub-pixel mapping.