期刊文献+
共找到4篇文章
< 1 >
每页显示 20 50 100
基于改进ResU-Net的中分辨遥感影像滑坡检测方法
1
作者 王颖 吴旭 +1 位作者 冷小鹏 余戈 《计算机技术与发展》 2023年第11期182-188,共7页
针对基于中分辨率遥感影像滑坡检测精度低的问题,提出了一种基于注意力机制的改进ResU-Net模型,并且基于多光谱遥感影像数据集得出了有益于滑坡检测的多特征模型输入组合。本研究所用的原始数据集共14个特征,首先剔除无效特征,并加入归... 针对基于中分辨率遥感影像滑坡检测精度低的问题,提出了一种基于注意力机制的改进ResU-Net模型,并且基于多光谱遥感影像数据集得出了有益于滑坡检测的多特征模型输入组合。本研究所用的原始数据集共14个特征,首先剔除无效特征,并加入归一化植被指数和归一化水体指数,生成新数据集。然后将新数据集应用于改进的ResU-Net与U-Net,ResU-Net,Attention U-Net,BiSeNet,Semantic FPN,U-Net++的对比实验,结果表明改进的ResU-Net在测试集上可获得76.91%的F1分数,同时精确率和召回率分别为77.34%和76.49%,在该任务中优于其他对比模型,且比ResU-Net模型的F1分数高了0.43百分点,有效提高了中分辨率遥感影像的滑坡检测精度。最后,再向数据集中依次加入归一化湿度指数和坡向特征,对比不同特征组合数据集产生的检测效果,结果发现加入坡向特征可最大化提升滑坡检测精度,F1分数可达77.03%。 展开更多
关键词 滑坡检测 多光谱 图像语义分割 注意力机制 ResU-Net
下载PDF
基于时间序列与时间卷积网络的滑坡位移预测 被引量:4
2
作者 江文金 冷小鹏 +2 位作者 林祥 冯梁玉 蒋浩 《科学技术与工程》 北大核心 2023年第9期3672-3679,共8页
滑坡位移预测作为滑坡监测预警的重要组成部分,对滑坡灾害的防治具有重要意义。目前,滑坡位移预测大多集中在循环架构的神经网络模型上,其存在梯度爆炸、消失问题等问题。为此,提出了一种基于时间序列与时间卷积网络(time convolution n... 滑坡位移预测作为滑坡监测预警的重要组成部分,对滑坡灾害的防治具有重要意义。目前,滑坡位移预测大多集中在循环架构的神经网络模型上,其存在梯度爆炸、消失问题等问题。为此,提出了一种基于时间序列与时间卷积网络(time convolution network,TCN)的滑坡位移预测模型。首先,该模型通过移动平均法将滑坡位移分解为趋势项位移和周期项位移。其次,采用Holt线性趋势模型预测趋势项位移,并建立时间卷积网络预测周期项位移。最后,将趋势项位移和周期项位移叠加,实现滑坡位移的预测。将该模型用于八字门滑坡的观测研究,结果表明:该模型相较于循环架构的神经网络模型能更有效地提取时序特征,预测精度更高。将基于TCN的滑坡位移预测模型应用于滑坡位移预测具有广阔的应用前景。 展开更多
关键词 滑坡位移预测 时间卷积网络 Holt线性趋势模型 八字门滑坡
下载PDF
Monitoring and Recognition of Debris Flow Infrasonic Signals 被引量:11
3
作者 LIU Dun-long leng xiao-peng +2 位作者 WEI Fang-qiang ZHANG Shao-jie HONG Yong 《Journal of Mountain Science》 SCIE CSCD 2015年第4期797-815,共19页
Low frequency infrasonic waves are emitted during the formation and movement of debris flows, which are detectable in a radius of several kilometers, thereby to serve as the precondition for their remote monitoring.Ho... Low frequency infrasonic waves are emitted during the formation and movement of debris flows, which are detectable in a radius of several kilometers, thereby to serve as the precondition for their remote monitoring.However, false message often arises from the simple mechanics of alarms under the ambient noise interference.To improve the accuracy of infrasound monitoring for early-warning against debris flows, it is necessary to analyze the monitor information to identify in them the infrasonic signals characteristic of debris flows.Therefore, a large amount of debris flow infrasound and ambient noises have been collected from different sources for analysis to sum up their frequency spectra, sound pressures, waveforms, time duration and other correlated characteristics so as to specify the key characteristic parameters for different sound sources in completing the development of the recognition system of debris flow infrasonic signals for identifying their possible existence in the monitor signals.The recognition performance of the system has been verified by simulating tests and long-term in-situ monitoring of debris flows in Jiangjia Gully,Dongchuan, China to be of high accuracy and applicability.The recognition system can provide the local government and residents with accurate precautionary information about debris flows in preparation for disaster mitigation and minimizing the loss of life and property. 展开更多
关键词 泥石流灾害 特征识别 监测预警 声信号 识别系统 次声波 运动过程 远程监控
下载PDF
Debris flows monitoring and localization using infrasonic signals 被引量:2
4
作者 leng xiao-peng LIU Dun-long +2 位作者 WEI Fang-qiang HONG Yong DAI De-fu 《Journal of Mountain Science》 SCIE CSCD 2017年第7期1279-1291,共13页
Infrasonic waves(frequency ≤ 20 Hz) are generated during the formation and movement of debris flows, traveling in air with a speed far higher than that of the debris-flow movement. Infrasound monitoring and localizat... Infrasonic waves(frequency ≤ 20 Hz) are generated during the formation and movement of debris flows, traveling in air with a speed far higher than that of the debris-flow movement. Infrasound monitoring and localization of infrasonic waves can serve as warning properties for debris-flows. Based on the characteristics of infrasonic signals, this study presents a three-point array of infrasound sensors as time-synchronous multiple sensors for acquiring signals. In the meantime, the signals are sorted by mutual correlation of signals to figure out their latency, and by means of array coordinating to locate the sound source to realize the monitoring and positioning of a debris-flows hazard. The method has been in situ tested and has been proven to be accurate in monitoring debris-flow occurrences and determining their positions, which is particularly effective for pre-event warning of debris-flow hazards. 展开更多
关键词 泥石流灾害 信号定位 监测定位 次声波 泥石流运动 运动过程 声波定位 声波信号
下载PDF
上一页 1 下一页 到第
使用帮助 返回顶部