Mixed attention mechanism on ResNet-DeepLabV3+ for
paddy field segmentation

Mixed attention mechanism on ResNet-DeepLabV3+ for
paddy field segmentation

Attention mechanism
DeepLabV3+
Remote sensing
Residual network
Semantic segmentation

Rice cultivation monitoring is crucial for Indonesia, where paddy field areas declined
by 2.45% according to the Central Bureau of Statistics due to land function
changes and shifting crop preferences. Regular monitoring of paddy field
distribution is essential for understanding agricultural land utilization by farmers
and landowners. Satellite imagery has become increasingly common for agricultural
land observation, but traditional neural networks alone provide insufficient
segmentation accuracy. This study proposes an enhanced deep learning architecture
combining residual network (ResNet)-DeepLabV3+ with coordinate attention
(CA) and spatial group-wise enhancement (SGE) modules. The attention
mechanisms establish direct connections between context vectors and inputs,
enabling the model to prioritize relevant spatial and spectral features for precise
paddy field identification. The CA module enhances spectral feature discrimination,
whereas the SGE improves spatial characteristic representation. The
experimental results demonstrate superior performance over the baseline methods,
achieving intersection over union (IoU) of 0.85, dice coefficient of 0.89,
and accuracy of 0.95. The proposed mixed attention mechanism significantly
improves the accuracy and efficiency of automatic crop area identification from
satellite imagery.

Alya Khairunnisa Rizkita1, Masagus Muhammad Luthfi Ramadhan1, Yohanes Fridolin Hestrio1,2,
Muhammad Hannan Hunafa1, Danang Surya Candra2, Wisnu Jatmiko1

Journal homepage: https://telkomnika.uad.ac.id/index.php/TELKOMNIKA

Sep 10, 2025

PERI IRAWAN

PDF

ENGLISH

TEXT