Identification of Stripe Rust and Leaf Rust on Different Wheat Varieties Based on Image Processing Technology

The timely and accurate identification of stripe rust leaf is essential in effective disease control the safe production wheat worldwide. To investigate methods for identifying two diseases on different varieties based image processing technology, single-leaf images varieties, acquired under field laboratory environmental conditions, were processed. After scaling, median filtering, morphological reconstruction, lesion segmentation images, 140 color, texture, shape features extracted from images; then, feature selections conducted using including ReliefF, 1R, correlation-based selection, principal components analysis combined with support vector machine (SVM), back propagation neural network (BPNN), random forest (RF), respectively. For individual-variety SVM, BPNN, RF models built optimal combinations, accuracies training sets testing same individual acquisition conditions as used modeling 87.18–100.00%, but most other low. multi-variety merged combinations range 82.05–100.00% achieved set, corresponding all sets. results indicated that could be greatly affected by satisfactory performances building multiple environments. This study provides an method a useful reference automatic plant diseases.