AUTHOR=Yang Tao , Wei Jingjing , Xiao Yongjun , Wang Shuyang , Tan Jingxuan , Niu Yupeng , Duan Xuliang , Pan Fei , Pu Haibo 

TITLE=LT-DeepLab: an improved DeepLabV3+ cross-scale segmentation algorithm for Zanthoxylum bungeanum Maxim leaf-trunk diseases in real-world environments

JOURNAL=Frontiers in Plant Science

VOLUME=Volume 15 - 2024

YEAR=2024

URL=https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2024.1423238

DOI=10.3389/fpls.2024.1423238

ISSN=1664-462X

ABSTRACT=Zanthoxylum bungeanum Maxim, an economically significant crop widespread in Asia, faces substantial yield declines due to frequent diseases in large-scale cultivation. Crop disease recognition using deep learning methods has become a prominent research area in agriculture. This paper introduces a novel model, LT-DeepLab, utilizing semantic segmentation to identify leaf spot(folium macula), rust, frost damage(gelu damnum), and diseased leaves and trunks in complex field environments. The proposed model enhances DeepLabV3+ with an innovative Fission Depth Separable with CRCC Atrous Spatial Pyramid Pooling module, which reduces the structural parameters of Atrous Spatial Pyramid Pooling module and improves cross-scale extraction capability. Incorporating Criss-Cross Attention with the Convolutional Block Attention Module provides a complementary boost to channel feature extraction. Additionally, deformable convolution enhances low-dimensional features, and a Fully Convolutional Network auxiliary header is integrated to optimize the network and enhance model accuracy without increasing parameter count. Compared to the baseline model, LT-DeepLab achieves a 3.59% improvement in mIoU, a 2.16% increase in mPA, and a 0.94% rise in OA, while reducing the computational demands by 11.11% and the parameter count by 16.82%.