Three-dimensional geological modeling is a process of interpreting geological features from limited sample data and making predictions, which can be converted into a classification task for grid units in the geological space. In sedimentary settings, it is difficult for a single geological classification process to comprehensively express the complex geological spatio-temporal relationships of underground space. In response to this problem, we proposed a progressive geological modeling strategy to reconstruct the subsurface based on a machine learning approach. The modeling work consisted of two-stage classifications. In the first stage, a stratigraphic classifier was built by mapping spatial coordinates into stratigraphic classes, which reflected the geological time information of the geological unit. Then, the obtained stratigraphic class was used as a new feature for the training of the lithologic classifier in the second stage, which allowed the stratigraphic information to be implicitly converted into a new rule condition and enabled us to output the lithologic class with stratigraphic implications. Finally, the joint Shannon entropy of two classifications was calculated to evaluate the uncertainty of the total steps. The experiment built a fine-grained 3D geological model with integrated expression of stratigraphic and lithologic information and validated the effectiveness of the strategy. Moreover, compared with the conventionally trained classifier, the misclassification of the lithologic class between different strata in the progressive classification results has been reduced, with the improvement of the F1-score from 0.75 to 0.78.

Imbalanced learning is a common problem in remote sensing imagery-based land-use and land-cover classifications. Imbalanced learning can lead to a reduction in classification accuracy and even the omission of the minority class. In this paper, an impartial semi-supervised learning strategy based on extreme gradient boosting (ISS-XGB) is proposed to classify very high resolution (VHR) images with imbalanced data. ISS-XGB solves multi-class classification by using several semi-supervised classifiers. It first employs multi-group unlabeled data to eliminate the imbalance of training samples and then utilizes gradient boosting-based regression to simulate the target classes with positive and unlabeled samples. In this study, experiments were conducted on eight study areas with different imbalanced situations. The results showed that ISS-XGB provided a comparable but more stable performance than most commonly used classification approaches (i.e., random forest (RF), XGB, multilayer perceptron (MLP), and support vector machine (SVM)), positive and unlabeled learning (PU-Learning) methods (PU-BP and PU-SVM), and typical synthetic sample-based imbalanced learning methods. Especially under extremely imbalanced situations, ISS-XGB can provide high accuracy for the minority class without losing overall performance (the average overall accuracy achieves 85.92%). The proposed strategy has great potential in solving the imbalanced classification problems in remote sensing.