研究目的

To demonstrate an automated way of crop disease identification on various leaf sample images corresponding to different crop species employing Local Binary Patterns (LBPs) for feature extraction and One Class Classification for classification.

研究成果

The developed model was trained on vine leaves to identify four different health conditions. The novelty of the current application is high generalization capability which was proven through testing in various leaf samples belonging to different plant species. The results proved that the model was efficient for most of the cases. More specifically, 44 of the 46 tested plant disease combination were successfully classified, giving a total success rate of 95%.

研究不足

Some limitations of the presented approach could possibly arise due to extrinsic factors such as image background and capture conditions and to intrinsic factors including segmentation and different disorders with similar symptoms.