Table 6 Comparative experiments of different data augmentation methods on the CIFAR-10 dataset.

From: Traditional cloud pattern classification algorithm based on semi-supervision with Random Line Augment

Data augmentation methods

Accuracy

Random horizontal flip

93.97%

Random cropping

93.97%

Random horizontal flip+Random crop

94.82%

Random horizontal flip+Random crop+RandAugment

93.97%

Random horizontal flip+Random crop+Roberts edge detection operator

92.23%

Random horizontal flip+Random crop+RandAugment+Roberts edge detection operator

93.10%