Classification-Based Urban Land Cover Prediction: A Machine Learning Method
Chiragh Goel
Apeejay School, Pitampura
Download PDFAbstract
Rapid urban expansion requires efficient monitoring and predicting changes in urban land cover. This study explores the application of machine learning (ML) classification algorithms to predict urban land cover. The study demonstrates how ML can accurately classify and predict land cover types in urban environments using publicly available datasets and advanced classification techniques. Therefore, the findings based on the suggested ML algorithms reveal that Random Forest (RF), Support Vector Machines (SVM), and Neural Networks (NN) can produce improved accuracy and scalabilities compared with traditional methods. This study highlights using preprocessing techniques such as normalization, feature extraction, and augmentation to enhance model performance. Comparative analyses of multiple algorithms help to identify the respective pros and cons of various algorithms. Random Forest is simple and efficient; however, Support Vector Machines can be robustly used in high-dimensional space. Convolutional Neural Networks are best fitted for automatically learning hierarchical features with maximum classification accuracy. Furthermore, the paper covers the role of spectral, spatial, and temporal features in enhancing predictive accuracy and evaluates the computational trade-offs of various approaches. The results indicate the potential of ML in urban planning and land-use management, providing scalable and accurate solutions for monitoring dynamic urban landscapes. By bridging the gap between traditional GIS-based methods and advanced machine learning, this study sets a foundation for future research integrating multi-source data and transfer learning techniques to tackle urbanization challenges effectively.
Keywords: Urban Land Cover; Machine Learning (ML) Classification; Random Forest (RF); Support Vector Machines (SVM); Neural Networks (NN); Urban Planning; Land-Use Management
- C. Small, “Global Analysis of Urban Reflectance,” Remote Sens. Environ., vol. 86, no. 1, pp. 99–115, 2003.
- P. Gamba and M. Aldrighi, “Urban Remote Sensing Using Multiple Endmember Spectral Mixture Models,” IEEE Trans. Geosci. Remote Sens., vol. 45, no. 6, pp. 1658–1675, 2007.
- G. Mountrakis, J. Im, and C. Ogole, “Support Vector Machines in Remote Sensing: A Review,” ISPRS J. Photogramm. Remote Sens., vol. 66, no. 3, pp. 247–259, 2011.
- J. R. Jensen et al., “Urban/Suburban Land Use Analysis Using Remote Sensing and GIS,” Photogramm. Eng. Remote Sens., vol. 63, no. 6, pp. 611–622, 2001.
- T. Gislason, J. Benediktsson, and J. Sveinsson, “Random Forests for Land Cover Classification,” Pattern Recognit. Lett., vol. 27, no. 4, pp. 294–300, 2006.
- X. Zhang et al., “Deep Learning for Land Use and Land Cover Classification,” IEEE Geosci. Remote Sens. Lett., vol. 14, no. 7, pp. 1028–1032, 2017.