Despite the significant impacts of biomass burning (BB) on global climate change and regional air pollution, there is a relative lack of research on the temporal trends and geographic patterns of BB in Northeast China (NEC). This study investigates the spatial–temporal distribution of BB and its impact on the atmospheric environment in the NEC region during 2004 to 2023 based on remote sensing satellite data and reanalyzed data, using the Siegel’s Repeated Median Estimator and Mann–Kendall test for trend analysis, HDBSCAN to identify significant BB change regions, and Moran’s Index to examine the spatial autocorrelation of BB. The obtained results indicate a fluctuating yet overall increasing BB trend, characterized by annual increases of 759 for fire point counts (FPC) and 12,000 MW for fire radiated power (FRP). BB predominantly occurs in the Songnen Plain (SNP), Sanjiang Plain (SJP), Liaohe Plain (LHP), and the transitional area between SNP and the adjacent Greater Khingan Mountains (GKM) and Lesser Khingan Mountains (LKM). Cropland and urban areas exhibit the highest growth in BB trends, each surpassing 60% (p < 0.05), with the most significant growth cluster spanning 68,634.9 km². Seasonal analysis shows that BB peaks in spring and autumn, with spring experiencing the highest severity. The most critical periods for BB are March–April and October–November, during which FPC and FRP contribute to over 80% of the annual total. This trend correlates with spring planting and autumn harvesting, where cropland FPC constitutes 71% of all land-cover types involved in BB. Comparative analysis of the aerosol extinction coefficient (AEC) between areas with increasing and decreasing BB indicates higher AEC in BB increasing regions, especially in spring, with the vertical transport of BB reaching up to 1.5 km. County-level spatial autocorrelation analysis indicates high–high clustering in the SNP and SJP, with a notable resurgence of autocorrelation in the SNP, suggesting the need for coordinated provincial prevention and control efforts. Finally, our analysis of the impact of BB on atmospheric pollutants shows that there is a correlation between FRP and pollutants, with correlations for PM_2.5, PM₁₀, and CO of 0.4, 0.4, and 0.5, respectively. In addition, the impacts of BB vary by region and season, with the most significant impacts occurring in the spring, especially in the SNP, which requires more attention. In summary, considering the escalating BB trend in NEC and its significant effect on air quality, this study highlights the urgent necessity for improved monitoring and strategic interventions.

With the rapid expansion of the low-altitude economy, detecting micro unmanned aerial vehicles (UAVs) in complex urban environments faces severe challenges. In optical surveillance, micro-UAVs typically appear as weak and small targets lacking distinct spatial semantics, rendering conventional appearance-based feature extraction highly ineffective. Furthermore, extreme lighting conditions and complex backgrounds easily submerge these subtle textures into noise. While event cameras can uniquely capture the high-frequency rotational characteristics of UAV rotors thanks to their microsecond temporal resolution, their sparse asynchronous outputs are highly susceptible to environmental clutter. To address this, we propose EGDNet, an event-guided frequency-aware dual-modal detection network optimized for robust UAV sensing. First, we developed a custom co-axial beam-splitting hardware platform to construct a strictly pixel-aligned dual-modal dataset of 43,090 frames, eliminating inter-modal parallax at the physical level. Second, our framework introduces a novel frequency-domain deconstruction pipeline utilizing a Temporal Fast Fourier Transform (T-FFT) to isolate the high-frequency mechanical rotor vibrations from low-frequency background clutter. Additionally, a Cross-Modal Synergistic Enhancement Module (CMSEM) and a Cross-Scale Adaptive Fusion Pyramid Network (CSAFPN) are designed to leverage event-derived high-frequency details for RGB feature restoration and to preserve sub-pixel target signatures during deep downsampling. Experimental results demonstrate that EGDNet achieves a mean average precision of 74.63% (AP@0.5) and an F1 score of 78.57%, significantly outperforming state-of-the-art fusion architectures. This work provides a robust visual perception solution for intelligent low-altitude surveillance.