利用FY-3D气象卫星的遥感监测数据,对邢台市2024年度夏季热岛效应进行了分析,结果显示:2024年夏季邢台市全市平均气温28.1℃,较常年同期偏高1.8℃,为显著偏高年份。其中6月、8月为显著偏高年份,7月为正常年份。邢台市强热岛区域主要集...利用FY-3D气象卫星的遥感监测数据,对邢台市2024年度夏季热岛效应进行了分析,结果显示:2024年夏季邢台市全市平均气温28.1℃,较常年同期偏高1.8℃,为显著偏高年份。其中6月、8月为显著偏高年份,7月为正常年份。邢台市强热岛区域主要集中在城镇化程度较高的地区以及钢厂等热源工厂集中地区,农业种植区域、山区以及植被覆盖茂密区域大多呈弱热岛或无热岛现象。陆表高温与下垫面关系密切,以不透水层为主的城区陆表温度较高。从空间分布上看,夏季发生热岛效应面积最大并且较强热岛级别以上的区域主要分布在市区,其次是城镇化程度相对较高的区域。从热岛强度等级分布来看,邢台市以无热岛区为主,主要分布在郊区耕地,强冷岛和较强冷岛主要分布在西部山区等植被覆盖比较茂密区域。强热岛区和较强热岛区面积较小,主要分布于人口密集的城区。Using the remote sensing monitoring data of FY-3D meteorological satellite, the heat island effect of Xingtai City in summer 2024 was carried out. The results show that the average temperature of Xingtai City in summer 2024 is 28.1˚C, which is 1.8˚C higher than the same period of normal years, which is a significantly higher year. Among them, June and August are significantly high years, and July is a normal year. The strong heat island area in Xingtai City is mainly concentrated in areas with a high degree of urbanization and areas where heat source factories such as steel mills are concentrated. Agricultural planting areas, mountainous areas, and areas with dense vegetation cover mostly show weak or no heat island phenomenon. The land surface temperature is closely related to the underlying surface, and the land surface temperature is higher in the urban area dominated by the impermeable layer. From the perspective of spatial distribution, the areas with the largest heat island effect in summer and above the strong heat island level are mainly distributed in urban areas, followed by areas with a relatively high degree of urbanization. According to the distribution of heat island intensity grade, the non-heat island area is the main area, mainly distributed in the suburban cultivated land, and the strong cold island and strong cold island are mainly distributed in the western mountainous area and other vegetation coverage areas. The area of strong heat island and stronger heat island is small, mainly distributed in densely populated urban areas.展开更多
为了解决飞机目标机动数据集缺失的问题,文章利用运动学建模生成了丰富的轨迹数据集,为网络训练提供了必要的数据支持。针对现阶段轨迹预测运动学模型建立困难及时序预测方法难以提取时空特征的问题,提出了一种结合Transformer编码器和...为了解决飞机目标机动数据集缺失的问题,文章利用运动学建模生成了丰富的轨迹数据集,为网络训练提供了必要的数据支持。针对现阶段轨迹预测运动学模型建立困难及时序预测方法难以提取时空特征的问题,提出了一种结合Transformer编码器和长短期记忆网络(Long Short Term Memory,LSTM)的飞机目标轨迹预测方法,即Transformer-Encoder-LSTM模型。新模型可同时提供LSTM和Transformer编码器模块的补充历史信息和基于注意力的信息表示,提高了模型能力。通过与一些经典神经网络模型进行对比分析,发现在数据集上,新方法的平均位移误差减小到0.22,显著优于CNN-LSTMAttention模型的0.35。相比其他网络,该算法能够提取复杂轨迹中的隐藏特征,在面对飞机连续转弯、大机动转弯的复杂轨迹时,能够保证模型的鲁棒性,提升了对于复杂轨迹预测的准确性。展开更多
林地在全球生态系统中扮演着至关重要的角色。但传统监督学习方法在林地提取上存在特征选择不精确与未能充分利用像元间的上下文关系等缺陷,导致林地提取精度不理想。针对上述问题,本文提出了一种基于改进PSPNet(Pyramid Scene Parsing ...林地在全球生态系统中扮演着至关重要的角色。但传统监督学习方法在林地提取上存在特征选择不精确与未能充分利用像元间的上下文关系等缺陷,导致林地提取精度不理想。针对上述问题,本文提出了一种基于改进PSPNet(Pyramid Scene Parsing Network)模型的高分辨率遥感影像林地提取方法。首先,利用高分二号遥感影像与全国第三次土地调查数据,制作高分辨率林地数据集。其次,通过在原始PSPNet模型的基础上加入SE(Squeeze and Excitation)注意力模块,改进PSPNet模型。实验结果表明,本文所改进的PSPNet模型的各项精度指标均优于其他方法,具有较高的提取精度。展开更多
文摘利用FY-3D气象卫星的遥感监测数据,对邢台市2024年度夏季热岛效应进行了分析,结果显示:2024年夏季邢台市全市平均气温28.1℃,较常年同期偏高1.8℃,为显著偏高年份。其中6月、8月为显著偏高年份,7月为正常年份。邢台市强热岛区域主要集中在城镇化程度较高的地区以及钢厂等热源工厂集中地区,农业种植区域、山区以及植被覆盖茂密区域大多呈弱热岛或无热岛现象。陆表高温与下垫面关系密切,以不透水层为主的城区陆表温度较高。从空间分布上看,夏季发生热岛效应面积最大并且较强热岛级别以上的区域主要分布在市区,其次是城镇化程度相对较高的区域。从热岛强度等级分布来看,邢台市以无热岛区为主,主要分布在郊区耕地,强冷岛和较强冷岛主要分布在西部山区等植被覆盖比较茂密区域。强热岛区和较强热岛区面积较小,主要分布于人口密集的城区。Using the remote sensing monitoring data of FY-3D meteorological satellite, the heat island effect of Xingtai City in summer 2024 was carried out. The results show that the average temperature of Xingtai City in summer 2024 is 28.1˚C, which is 1.8˚C higher than the same period of normal years, which is a significantly higher year. Among them, June and August are significantly high years, and July is a normal year. The strong heat island area in Xingtai City is mainly concentrated in areas with a high degree of urbanization and areas where heat source factories such as steel mills are concentrated. Agricultural planting areas, mountainous areas, and areas with dense vegetation cover mostly show weak or no heat island phenomenon. The land surface temperature is closely related to the underlying surface, and the land surface temperature is higher in the urban area dominated by the impermeable layer. From the perspective of spatial distribution, the areas with the largest heat island effect in summer and above the strong heat island level are mainly distributed in urban areas, followed by areas with a relatively high degree of urbanization. According to the distribution of heat island intensity grade, the non-heat island area is the main area, mainly distributed in the suburban cultivated land, and the strong cold island and strong cold island are mainly distributed in the western mountainous area and other vegetation coverage areas. The area of strong heat island and stronger heat island is small, mainly distributed in densely populated urban areas.
文摘为了解决飞机目标机动数据集缺失的问题,文章利用运动学建模生成了丰富的轨迹数据集,为网络训练提供了必要的数据支持。针对现阶段轨迹预测运动学模型建立困难及时序预测方法难以提取时空特征的问题,提出了一种结合Transformer编码器和长短期记忆网络(Long Short Term Memory,LSTM)的飞机目标轨迹预测方法,即Transformer-Encoder-LSTM模型。新模型可同时提供LSTM和Transformer编码器模块的补充历史信息和基于注意力的信息表示,提高了模型能力。通过与一些经典神经网络模型进行对比分析,发现在数据集上,新方法的平均位移误差减小到0.22,显著优于CNN-LSTMAttention模型的0.35。相比其他网络,该算法能够提取复杂轨迹中的隐藏特征,在面对飞机连续转弯、大机动转弯的复杂轨迹时,能够保证模型的鲁棒性,提升了对于复杂轨迹预测的准确性。
文摘林地在全球生态系统中扮演着至关重要的角色。但传统监督学习方法在林地提取上存在特征选择不精确与未能充分利用像元间的上下文关系等缺陷,导致林地提取精度不理想。针对上述问题,本文提出了一种基于改进PSPNet(Pyramid Scene Parsing Network)模型的高分辨率遥感影像林地提取方法。首先,利用高分二号遥感影像与全国第三次土地调查数据,制作高分辨率林地数据集。其次,通过在原始PSPNet模型的基础上加入SE(Squeeze and Excitation)注意力模块,改进PSPNet模型。实验结果表明,本文所改进的PSPNet模型的各项精度指标均优于其他方法,具有较高的提取精度。
