Interpretable Deep Learning Framework for Land Use and Land Cover Classification in Remote Sensing Using SHAP

• Published in: IEEE geoscience and remote sensing letters Vol. 20; pp. 1 - 5
• Main Authors: Temenos, Anastasios, Temenos, Nikos, Kaselimi, Maria, Doulamis, Anastasios, Doulamis, Nikolaos
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 2023; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Accuracy; Additives; Artificial neural networks; Classification; Convolutional neural network (CNN); Convolutional neural networks; Correlation; Crops; Datasets; Deep learning; EuroSAT; explainable AI (XAI); Frameworks; Image classification; Land cover; Land use; Machine learning; Neural networks; Predictions; Remote sensing; Rural areas; Satellite imagery; Shapley additive explanation (SHAP)
• ISSN: 1545-598X; 1558-0571
• DOI: 10.1109/LGRS.2023.3251652

An interpretable deep learning framework for land use and land cover (LULC) classification in remote sensing using Shapley additive explanations (SHAPs) is introduced. It utilizes a compact convolutional neural network (CNN) model for the classification of satellite images and then feeds the results to a SHAP deep explainer so as to strengthen the classification results. The proposed framework is applied to Sentinel-2 satellite images containing 27000 images of pixel size <inline-formula> <tex-math notation="LaTeX">64 \times 64 </tex-math></inline-formula> and operates on three-band combinations, reducing the model's input data by 77% considering that 13 channels are available, while at the same time investigating on how different spectrum bands affect predictions on the dataset's classes. Experimental results on the EuroSAT dataset demonstrate the CNN's accurate classification with an overall accuracy of 94.72 %, whereas the classification accuracy on three-band combinations on each of the dataset's classes highlights its improvement when compared to standard approaches with larger number of trainable parameters. The SHAP explainable results of the proposed framework shield the network's predictions by showing correlation values that are relevant to the predicted class, thereby improving the classifications occurring in urban and rural areas with different land uses in the same scene.