In this study, a convolutional neural network (CNN) is developed that predicts crop type in the US great plains using optical and active microwave remote sensing, historical crop type data and economic commodity crop price information. It includes a Markov chain model that predicts the prior probability of crop types from historical crop type sequences for each pixel. This prior and the remotely sensed imagery is fed through a CNN to produce the final crop type. The CNN is trained using imagery and crop type from the previous year. The field-scale accuracy of this method is found to be 8̃5% using imagery from just the first 90 growing days.

In this study, a machine learning based algorithm is developed for dynamically classifying and determining area of agricultural land-use using optical and active microwave remote sensing. It includes a gradient boosted machine trained in the US using labels derived from the United States Department of Agriculture National Agricultural Statistics Service crop land data layers (CDL). This is then applied to the state of Mato Grosso in Brazil with a hidden Markov model applied to temporally stabilize the land-cover predictions. High resolution remote sensing products such as land surface temperature, normalized difference vegetation index, nadir adjusted bidirectional reflectance, land-cover and radar backscatter were used to develop the classification model. The results in the US were validated using a hold-out set with CDL labels. The results in Brazil were validated using the harvested area statistics reported by Companhia Nacional de Abastecimento.

Active Label Correction (ALC) is an interactive method that cleans an established training set of mislabeled examples in conjunction with a domain expert. ALC presumes that the expert who conducts this review is either more accurate than the original annotator or has access to additional resources that ensure a high quality label. A high-cost re-review is possible because ALC proceeds iteratively, scoring the full training set but selecting only small batches of examples that are likely mislabeled. The expert reviews each batch and corrects any mislabeled examples, after which the classifier is retrained and the process repeats until the expert terminates it. We compare several instantiations of ALC to fully-automated methods that attempt to discard or correct label noise in a single pass. Our empirical results show that ALC outperforms single-pass methods in terms of selection efficiency and classifier accuracy. We evaluate the best ALC instantiation on our motivating task of detecting mislabeled and poorly formulated sites within a land cover classification training set from the geography domain.

Two aspects are crucial when constructing any real world supervised classification task: the set of classes whose distinction might be useful for the domain expert, and the set of classifications that can actually be distinguished by the data. Often a set of labels is defined with some initial intuition but these are not the best match for the task. For example, labels have been assigned for land cover classification of the Earth but it has been suspected that these labels are not ideal and some classes may be best split into subclasses whereas others should be merged. This paper formalizes this problem using three ingredients: the existing class labels, the underlying separability in the data, and a special type of input from the domain expert. We require a domain expert to specify an L × L matrix of pairwise probabilistic constraints expressing their beliefs as to whether the L classes should be kept separate, merged, or split. This type of input is intuitive and easy for experts to supply. We then show that the problem can be solved by casting it as an instance of penalized probabilistic clustering (PPC). Our method, Class-Level PPC (CPPC) extends PPC showing how its time complexity can be reduced from O(N2) to O(NL) for the problem of class re-definition. We further extend the algorithm by presenting a heuristic to measure adherence to constraints, and providing a criterion for determining the model complexity (number of classes) for constraint-based clustering. We demonstrate and evaluate CPPC on artificial data and on our motivating domain of land cover classification. For the latter, an evaluation by domain experts shows that the algorithm discovers novel class definitions that are better suited to land cover classification than the original set of labels.