Parcel-Based Active Learning for Large Extent Cultivated Area Mapping

This paper focuses on agricultural land cover mapping at a high-resolution scale and over large areas from an operational point of view and from a high-resolution monodate image. In this context, training data are assumed to be collected by successive journeys of field surveys and, thus, are very limited. Supervised learning techniques are generally used, assuming that the classes distribution is constant over the whole image. However, in practice, a data shift often occurs on large areas due to various acquisition conditions. To alleviate these issues, active learning (AL) techniques define an efficient training set by iteratively adapting it through adding the most informative unlabeled instances. They can improve the classification process efficiency while keeping a limited training dataset. The novelty in this paper is the application of AL techniques on multispectral images for agricultural land cover mapping, using field sampling instead of pixel sampling, which is rarely done in the literature. Besides, we proposed a parcel-based AL scheme that is suitable for an operational land cover mapping in cultivated areas since the parcel is an agricultural unit and field observations are processed at parcel scale. Random forests classifier was used. Results were processed on a 6 m multispectral Spot6 image over a 35 km $^2$ Mediterranean cultivated area, in Lebna Catchment, north eastern Tunisia. The contribution of AL techniques was assessed with comparison to a random and stratified random strategies for sampling new instances. For iterative sample selection, two criteria are used and often coupled: uncertainty and diversity. For diversity metric, a new clustering-based metric was proposed based on a mean-shift clustering, which improved the classification accuracy. AL techniques showed to be efficient with complex data and fine land cover legend improving random-based selection up to 10%. Besides, the maximum of classification accuracy is reached using mean-shift breaking ties metric in just 5-day field survey, i.e., 30 days less compared to the random selection. Finally, results showed that the finer the definition of land cover classes, the more crucial is the choice of AL metrics.

[1]  Lorenzo Bruzzone,et al.  Active and Semisupervised Learning for the Classification of Remote Sensing Images , 2014, IEEE Transactions on Geoscience and Remote Sensing.

[2]  Heejun Chang,et al.  Effects of land cover, topography, and built structure on seasonal water quality at multiple spatial scales. , 2012, Journal of hazardous materials.

[3]  Gérard Dedieu,et al.  Assessing the robustness of Random Forests to map land cover with high resolution satellite image time series over large areas , 2016 .

[4]  Melba M. Crawford,et al.  Multi-view adaptive disagreement based active learning for hyperspectral image classification , 2010, 2010 IEEE International Geoscience and Remote Sensing Symposium.

[5]  Antonio J. Plaza,et al.  Superpixel-Based Active Learning and Online Feature Importance Learning for Hyperspectral Image Analysis , 2017, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.

[6]  Daphne Koller,et al.  Support Vector Machine Active Learning with Applications to Text Classification , 2000, J. Mach. Learn. Res..

[7]  Lorenzo Bruzzone,et al.  An Effective Strategy to Reduce the Labeling Cost in the Definition of Training Sets by Active Learning , 2014, IEEE Geoscience and Remote Sensing Letters.

[8]  William J. Emery,et al.  Active Learning Methods for Remote Sensing Image Classification , 2009, IEEE Transactions on Geoscience and Remote Sensing.

[9]  Shlomo Argamon,et al.  Committee-Based Sampling For Training Probabilistic Classi(cid:12)ers , 1995 .

[10]  Farid Melgani,et al.  Model-based active learning for SVM classification of remote sensing images , 2010, 2010 IEEE International Geoscience and Remote Sensing Symposium.

[11]  Joydeep Ghosh,et al.  An Active Learning Approach to Hyperspectral Data Classification , 2008, IEEE Transactions on Geoscience and Remote Sensing.

[12]  Mikhail F. Kanevski,et al.  A Survey of Active Learning Algorithms for Supervised Remote Sensing Image Classification , 2011, IEEE Journal of Selected Topics in Signal Processing.

[13]  David A. Cohn,et al.  Improving generalization with active learning , 1994, Machine Learning.

[14]  Marin Ferecatu,et al.  Interactive Remote-Sensing Image Retrieval Using Active Relevance Feedback , 2007, IEEE Transactions on Geoscience and Remote Sensing.

[15]  Lorenzo Bruzzone,et al.  Definition of Effective Training Sets for Supervised Classification of Remote Sensing Images by a Novel Cost-Sensitive Active Learning Method , 2014, IEEE Transactions on Geoscience and Remote Sensing.

[16]  P. Lagacherie,et al.  Fuzzy k-means clustering of fields in an elementary catchment and extrapolation to a larger area , 1997 .

[17]  Lorenzo Bruzzone,et al.  Batch-Mode Active-Learning Methods for the Interactive Classification of Remote Sensing Images , 2011, IEEE Transactions on Geoscience and Remote Sensing.

[18]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[19]  Saiful Islam,et al.  Mahalanobis Distance , 2009, Encyclopedia of Biometrics.

[20]  Sankar K. Pal,et al.  Segmentation of multispectral remote sensing images using active support vector machines , 2004, Pattern Recognit. Lett..

[21]  Roger Moussa,et al.  Impact of the spatial arrangement of land management practices on surface runoff for small catchments , 2012 .

[22]  Mahesh Pal,et al.  Random forest classifier for remote sensing classification , 2005 .

[23]  B. Amiri,et al.  Modeling the Linkage Between River Water Quality and Landscape Metrics in the Chugoku District of Japan , 2009 .

[24]  Leo Breiman,et al.  Bagging Predictors , 1996, Machine Learning.

[25]  Ion Muslea,et al.  Active Learning with Multiple Views , 2009, Encyclopedia of Data Warehousing and Mining.

[26]  Mark A. Girolami,et al.  Mercer kernel-based clustering in feature space , 2002, IEEE Trans. Neural Networks.

[27]  Yizong Cheng,et al.  Mean Shift, Mode Seeking, and Clustering , 1995, IEEE Trans. Pattern Anal. Mach. Intell..

[28]  Dorin Comaniciu,et al.  Mean Shift: A Robust Approach Toward Feature Space Analysis , 2002, IEEE Trans. Pattern Anal. Mach. Intell..

[29]  Andrew McCallum,et al.  Employing EM and Pool-Based Active Learning for Text Classification , 1998, ICML.