Abstract:Accurate remote sensing-based land use and crop maps provide important and timely information for decision support in large-scale agricultural monitoring. Most existing multi-crop products for complex agricultural areas based on traditional machine learning algorithms are not suitable for large agricultural management because of the poor model transfer capabilities. Therefore, a deep segmentation and classification model including spatiotemporal transfer across regions and years must be developed. In this study, a deep learning approach was developed based on the UNet++ architecture by integrating feature fusion and upsampling of small samples for large-scale land use and crop mapping. Classification experiments were conducted for ten categories at four sites using 10 m resolution Sentinel-2A images from 2019 to 2021 containing 4,194,304 pixels. The joint loss, including the label smoothing cross entropy and Dice coefficient, and the mean intersection over union (mIoU) were used to evaluate the model performance depending on different features and upsampling schemes. The joint loss and mIoU values of the training model and the prediction accuracies of the test sites indicate that the scheme including the upsampling and fusion of spectral bands, vegetation indices, and texture features yields the optimal model performance. For comparison, UNet, DeepLab V3+, Pyramid Scene Parsing Network (PSPNet), and random forest models were built. The improved UNet++ model exhibits the best performance, with an overall accuracy, kappa, and macro F1 above 91 %, 85 %, and 51 %, respectively. The deep segmentation and classification results without training samples demonstrate the spatiotemporal transfer capability of the UNet++ architecture during the key crop growth period. The patch parameter values also indicate that the improved model exhibits a better shape compactness and significantly reduces and suppresses patch fragmentation. Based on the analysis of the overall confusion matrix, the proposed method improves the classification accuracy for datasets with imbalanced land cover and crop types. This study provides a strategy for large-scale and complex land use and crop mapping based on the integration of feature fusion and deep learning.

Weighted skip-connection feature fusion: A method for augmenting UAV oriented rice panicle image segmentation.

An Image Segmentation Method Based on Transformer and Multi-Scale Feature Fusion for UAV Marine Environment Monitoring

Real-time Water Area Segmentation for USV using Enhanced U-Net

Lightweight Deep Learning Models for High-Precision Rice Seedling Segmentation from UAV-Based Multispectral Images

FG-UNet: fine-grained feature-guided UNet for segmentation of weeds and crops in UAV images

Multi-scale Feature Extraction and Fusion Net: Research on UAVs Image Semantic Segmentation Technology

UAM-Net: an Attention-Based Multi-level Feature Fusion UNet for Remote Sensing Image Segmentation.

Semantic segmentation of remote sensing images combined with attention mechanism and feature enhancement U-Net

A deep learning method for optimizing semantic segmentation accuracy of remote sensing images based on improved UNet

Crop Segmentation of Unmanned Aerial Vehicle Imagery Using Edge Enhancement Network

Refined Feature Fusion for In-Field High-Density and Multi-Scale Rice Panicle Counting in UAV Images

Research on the Classification of Complex Wheat Fields Based on Multi-Scale Feature Fusion

UAVSeg: Dual-Encoder Cross-Scale Attention Network for UAV Images Semantic Segmentation

Rice Counting and Localization in Unmanned Aerial Vehicle Imagery Using Enhanced Feature Fusion

Real-Time Identification of Rice Weeds by UAV Low-Altitude Remote Sensing Based on Improved Semantic Segmentation Model

Link Aggregation for Skip Connection–Mamba: Remote Sensing Image Segmentation Network Based on Link Aggregation Mamba

Deep Segmentation and Classification of Complex Crops Using Multi-Feature Satellite Imagery

HSI-TransUNet: A transformer based semantic segmentation model for crop mapping from UAV hyperspectral imagery

Iae-Sdnet: an End-to-End Image Adaptive Enhancement and Wheat Scab Detection Network Using Uav

Transferring learned patterns from ground-based field imagery to predict UAV-based imagery for crop and weed semantic segmentation in precision crop farming

Adaptively Spatial Feature Fusion Network: an Improved UAV Detection Method for Wheat Scab