Change detection is crucial for evaluating land use, land cover changes, and sustainable development, constituting a significant component of Earth observation tasks. The difficulty in extracting features from high-resolution images, coupled with the complexity of image content, poses challenges for traditional change detection algorithms in terms of accuracy and applicability. The recent emergence of deep learning methods has led to substantial progress in the field of change detection. However, existing frameworks often involve the simplistic integration of bi-temporal features in specific areas, lacking the fusion of temporal information and semantic details in the images. In this paper, we propose an attention-guided multi-scale fusion network (AMFNet), which effectively integrates bi-temporal image features and diverse semantics at both the encoding and decoding stages. AMFNet utilizes a unique attention-guided mechanism to dynamically adjust feature fusion, enhancing adaptability and accuracy in change detection tasks. Our method intelligently incorporates temporal information into the deep learning model, considering the temporal dependency inherent in these tasks. We decode based on an interactive feature map, which improves the model’s understanding of evolving patterns over time. Additionally, we introduce multi-level supervised training to facilitate the learning of fused features across multiple scales. In comparison with different algorithms, our proposed method achieves F1 values of 0.9079, 0.8225, and 0.8809 in the LEVIR-CD, GZ-CD, and SYSU-CD datasets, respectively. Our model outperforms the SOTA model, SAGNet, by 0.69% in terms of F1 and 1.15% in terms of IoU on the LEVIR-CD dataset, by 2.8% in terms of F1 and 1.79% in terms of IoU on the GZ-CD dataset, and by 0.54% in terms of F1 and 0.38% in terms of IoU on the SYSU-CD dataset. The method proposed in this study can be applied to various complex scenarios, establishing a change detection method with strong model generalization capabilities.

Rubber (Hevea brasiliensis Muell.) plantations are among the most critical agricultural ecosystems in tropical regions, playing a vital role in regional carbon balance. Accurate large-scale biomass estimation for these plantations remains a challenging task due to the severe signal saturation problem. Recent advances in remote sensing big data, cloud platforms, and machine learning have facilitated the precise acquisition of key physiological variables, such as stand age (A) and canopy height (H), which are critical parameters for biomass estimation but have been underutilized in prior studies. Using Hainan Island—the second-largest rubber planting base in China—as a case study, we integrated extensive ground surveys, maps of stand age and canopy height, remote sensing indicators (RSIs), and geographical and climate indicators (ECIs) to ascertain the optimal method for estimating rubber plantation biomass. We compared different inputs and estimation approaches (direct and indirect) using the random forest algorithm and analyzed the spatiotemporal characteristics of rubber plantation biomass on Hainan Island. The results indicated that the traditional model (RSIs + ECIs) had low accuracy and significant estimation bias (R² = 0.24, RMSE = 38.36 mg/ha). The addition of either stand age or canopy height considerably enhance model accuracy (R² = 0.77, RMSE ≈ 21.12 mg/ha). Moreover, incorporating the DBH obtained through indirect inversion yielded even greater predictive accuracy (R² = 0.97, RMSE = 7.73 mg/ha), outperforming estimates derived from an allometric equation model input with the DBH (R² = 0.67, RMSE = 25.43 mg/ha). However, augmenting the model with stand age, canopy height, or their combination based on RSIs, ECIs, and DBH only marginally improved the accuracy. Consequently, it is not recommended in scenarios with limited data and computing resources. Employing the optimal model, we generated biomass maps of rubber plantations on Hainan Island for 2016 and 2020, revealing that the spatiotemporal distribution pattern of the biomass is closely associated with the establishment year of the rubber plantations. While average biomass in a few areas has undergone slight decreases, total biomass has exhibited significant growth, reaching 5.46 × 10⁷ mg by the end of 2020, underscoring its considerable value as a carbon sink.