Rice Blast Disease Identification Based on RCM-Enhanced DeeplabV3+ Model

Rice blast is one of the most harmful diseases to the growth of rice, which can cause a 10 % to 30 % reduction in rice production each year and even lead to a complete crop failure in severe cases. Therefore, accurately identifying rice blast would help to improve rice production and reduce losses. In the end, this paper proposes an RCM-Enhanced DeeplabV3+ model. The model incorporates a new multi-scale feature extraction network and a feature enhancement module, which pay more attention to the axial context information of the target area, thereby improving the feature extraction capabilities of small area lesions. Additionally, the three-leaf indica rice variety CO39 was selected as the research subject, and a set of 722 images of rice leaf diseases was constructed for simulation experiments. The associated experimental results show that the RCM-Enhanced DeeplabV3+ model achieves <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$82.31 \%, 90.33 \%$</tex>, and 89.49 % in mIoU, mPA and F1- Macro coefffcients, respectively, outperforming the comparative methods.