More convolutional neural network (CNN) methods are widely utilized for ship detection in synthetic aperture radar (SAR) images. Nevertheless, there are still some problems that need to be addressed. First, detection methods may misclassify islands and onshore objects as ships. Second, due to the large scale and low resolution, a majority of ships are regarded as small targets and difficult to detect. Third, due to the different sizes of ships, the hyperparameters of anchor-based methods are difficult to set. Thus, we propose an anchor-free method called improved fully convolutional one-stage object detector (Improved-FCOS) to solve these problems. First, a multilevel feature attention mechanism is proposed to extract effective features and collect global context information. Second, we propose a feature refinement and reuse module with two stages to refine small ship features. Third, a head improvement module is designed to optimize the methods for classification and localization of ship targets. Finally, a modified varifocal loss is adopted to better train the classification branch. We conduct ablation experiments on LS-SSDD-v1.0 and verify the detection performance of the proposed modules. Additionally, we compare the Improved-FCOS with other state-of-the-art methods by experiments on three datasets, including the SAR-Ship-Dataset, HRSID, and LS-SSDD-v1.0. Detection results show that Improved-FCOS obtains the best detection performance on all datasets and indicate that the Improved-FCOS is more accurate and robust.
23 August 2023 in IEEE Geoscience and Remote Sensing Letters
Ship detection in synthetic aperture radar (SAR) images using a convolutional neural network (CNN) has become a hotspot. However, dense multi-size ship targets in inshore scenes, proximity of ship targets to man-made targets, and datasets containing only one class limit the performance of ship detection methods. In pursuit of attaining heightened performance, most existing methods endeavor to augment the number of parameters and enhance network complexity. To address these problems, this letter proposes an anchor-free ghost feature extraction and cross-scale interaction network (GFECSI-Net), which improves detection performance through an efficient implementation, thereby avoiding the increase in the number of parameters or network complexity. First, to enhance the capability of feature extraction for ship targets, a multi-scale adaptive feature pyramid network (MSAFPN) is proposed to realize intensive information interaction and cross-scale feature fusion between different feature maps. Meanwhile, a selective efficient channel attention module (SECAM) is designed to enable the network to prioritize channels that better characterize ship targets. Besides, a GPU-efficient backbone for generating ghost feature maps and a task alignment detection head are integrated into GFECSI-Net. Comparison results with nine state-of-the-art CNN methods on a high-resolution SAR image dataset (HRSID) and a large-scale multi-class SAR target dataset (MSAR-1.0) indicate that GFECSI-Net achieves superior detection performance in F1-Score and mAP with a small number of parameters.
24 April 2025 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
The spatiotemporal prediction of RZSM refers to the process of estimating its future spatial distribution and temporal variations using predictive models. Accurate spatiotemporal predictions of soil moisture provide insights into future conditions, supporting decision-making in applications such as crop yield optimization, irrigation planning, and drought management. However, existing models face limitations in capturing complex spatiotemporal dependencies and dynamic causal interactions. This paper proposes a spatiotemporal prediction framework that integrates causal inference with deep learning, termed the Causal-Guided Spatiotemporal Swin Transformer (Causal ST-SwinT). The model introduces a dynamic causal weight adjustment mechanism to adaptively optimize the causal relationship intensity between variables and adopts a hierarchical multi-level feature extraction strategy to effectively capture complex spatiotemporal dependencies, thereby enhancing prediction accuracy and model interpretability. The proposed method is validated on the ERA5 and Soil Moisture Active Passive (SMAP) datasets over the Tibetan Plateau and compared with multiple models. Experimental results show that Causal ST-SwinT significantly outperforms the classical Convolutional Long Short-Term Memory (ConvLSTM) model, reducing mean absolute error (MAE) from 0.0146 to 0.0055 m³/m³ on the ERA5 dataset and from 0.0088 to 0.0046 m³/m³ on the SMAP dataset. Robustness analysis reveals that Causal ST-SwinT maintains high prediction accuracy under various environmental conditions. Ablation experiments further confirm the critical role of the causal attention module in improving model performance. These findings demonstrate that integrating causal knowledge with deep learning models effectively enhances the modeling capabilities of complex spatiotemporal systems, providing a novel solution for broader spatiotemporal prediction tasks.
27 November 2025 in Geo-spatial Information Science
Effective management of rye cover crops in cash-crop systems relies heavily on accurate biomass estimation. Low-altitude Unmanned Aerial Vehicle (UAV) imagery offers a promising high-resolution alternative, yet unlocking its full potential requires moving beyond basic estimation models to more integrative and explanatory models. This study obtains the measured height (MH), SPAD (Soil and Plant Analyzer Development) values, and measured dry biomass (MDB) and applies UAV remote sensing and machine learning to acquire the crop canopy height, vegetation indices (VIs), and vegetation fraction (VF) across growth stages. Among single-parameter biomass estimation models, the estimated height yields the best at the overall growth stage (R2 = 0.935), whereas selected VIs perform the best at the non-seedling stage (R2 = 0.851). For multi-parameters modeling, models combining height, VF, and VIs significantly outperform the single-parameter models, achieving better estimation results throughout each growth stage (Best R2 = 0.951). Structural equation modeling clarifies the direct and indirect contributions of these parameters to biomass accumulation, revealing their synergistic effects. This study demonstrates the potential of UAV-based multi-parameter biomass estimation model to support more informed decisions in cover crop management and to advance broader precise agriculture practices. Additionally, the analytical framework developed here offers a transferable approach for high-resolution biomass monitoring in other crop systems.
Underwater image enhancement is essential for marine visual perception tasks. However, the highly heterogeneous optical degradations in real-world waters, the scarcity of paired training data, and the inherent dilemma for existing models in balancing long-range dependency modeling with computational overhead pose significant challenges. To address these issues, this paper proposes a prior-guided, self-supervised underwater image enhancement framework called Scene-Aware Retinex Mamba (SARM). This framework seamlessly integrates Retinex theoretical priors with state space models (SSMs) and operates without paired supervision by employing a prior-guided pseudo-labeling strategy to guide network optimization. Architecturally, SARM deeply couples the physical Retinex prior with SSM. Its core module integrates multi-color space features and leverages a 2D selective scan mechanism to achieve global context modeling with linear complexity , effectively removing complex color casts and suppressing non-uniform scattering noise. To further overcome the generalization bottlenecks in cross-domain underwater testing, this paper introduces a Scene-Aware Adapter (SAA), which facilitates dynamic loss scheduling and adaptive feature gating by quantifying scene-specific degradation characteristics. Comprehensive evaluations on multiple benchmark datasets, including UIEB, EUVP, and UCCS, demonstrate that SARM achieves state-of-the-art subjective and objective enhancement quality (e.g., yielding a URanker score of 2.491 and a CCF score of 35.76), while maintaining an ultra-fast inference speed of 136.52 FPS on the UIEB dataset. Furthermore, extended experiments reveal that SARM can significantly boost the performance of downstream vision tasks, validating its potential as a robust preprocessing module for various practical marine vision applications.