Fig. 13

(a) Learning curve of the classifier using the multispectral satellite imagery alone, the vertical axis shows the accuracy (F1-score), and the horizontal axis represents the number of training transects. (b) Importance of features (mean decrease in impurity method) used in this classifier. The bars represent the average of the importance of each feature from multispectral satellite imagery (blue). The black line indicates the standard deviation of the importance of each feature.