| 18 | 0 | 21 |
| 下载次数 | 被引频次 | 阅读次数 |
检测电网工程建设实施过程中产生的水土扰动威胁，对维护当地生态稳定具有重要意义。该文针对电网施工建设带来的水土扰动检测问题进行了研究，利用高分辨率遥感影像构建了一套水土扰动变化检测数据集；结合多尺度差异特征融合与注意力机制，设计了MPANet变化检测算法用于水土扰动异常区自动识别；并改进Copy-Paste算法，针对变化区域添加对象级的尺度扰动，增加正样本比例。实验结果表明，MPANet变化检测算法在水土扰动变化检测数据集中的Precision、Recall、F1-score精度分别达到了84.12%、82.75%、83.43%,其结果充分证明了该文算法与数据集具备进行水土扰动异常区域检测的可行性，为电网工程建设过程中水土保持自动化监测与综合评估提供了技术参考。
Abstract：Detecting the threats of soil and water disturbance generated during the implementation of power grid construction is of great significance for preserving local ecological stability. This paper addresses the issue of monitoring soil and water disturbance caused by power grid construction, constructs a dataset for detecting changes in soil and water disturbances using high-resolution remote sensing images. Subsequently, by integrating multi-scale differential feature fusion and attention mechanism, we design a MPANet change detection algorithm for the automatic identification of water and soil disturbance anomaly areas. Furthermore, we improve the Copy-Paste algorithm focusing on object-level scale disturbances in the changed areas by increasing the proportion of positive samples. Experimental results show that the precision of Precision, Recall, and F1-score of this algorithm in the soil and water disturbance change detection dataset reach 84.12%, 82.75%, and 83.43% respectively.This fully proves the feasibility of the algorithm and dataset in this paper for water and soil disturbance anomaly detection, providing technical reference for the automated monitoring and comprehensive assessment of water and soil conservation throughout the entire process of power grid construction.
[1] 戚国，辉洪倩，陈枫楠，等.特高压塔基施工扰动区域快速自动识别研究.人民黄河，2019,41(08):99～102.
[2] 彭向阳，钱金菊，吴功平，等.架空输电线路机器人全自主巡检系统及示范应用.高电压技术，2017,43(8):2582～2591.
[3] 付晶，谈家英，周立玮，等.输电线路无人机巡检图像缺陷智能识别方法分析.高电压技术，2023,49(增刊1):103～110.
[4] 李德仁，丁霖，邵振峰.面向实时应用的遥感服务技术.遥感学报，2021,25(1):15～24.
[5] Malila W A.Change vector analysis:An approach for detecting forest changes with Landsat.1980:385.
[6] Celik T.Unsupervised change detection in satellite images using principal component analysis and k-means clustering.IEEE Geoscience and Remote Sensing Letters,2009,6(4):772～776.
[7] Feng Y,Jiang J,Xu H,et al.Change detection on remote sensing images using dual-branch multilevel intertemporal network.IEEE Transactions on Geoscience and Remote Sensing,2023,61:1～15.
[8] 杨彬，毛银，陈晋，等.深度学习的遥感变化检测综述：文献计量与分析.遥感学报，2023,27(9):1988～2005.
[9] 陈智朗，付振华，朱紫阳，等.基于HRNet的高分辨率遥感影像建筑物变化信息提取.测绘通报，2022(5):126～132.
[10] Chen H,Shi Z.A spatial-temporal attention-based method and a new dataset for remote sensing image change detection.Remote Sensing,2020,12(10):1662.
[11] Fang S,Li K,Li Z.Changer:Feature interaction is what you need for change detection.IEEE Transactions on Geoscience and Remote Sensing,2023,61:1～11.
[12] Fu J,Liu J,Jiang J,et al.Scene segmentation with dual relation-aware attention network.IEEE Transactions on Neural Networks and Learning Systems,2021,32(6):2547～2560.
[13] Ghiasi G,Cui Y,Srinivas A,et al.Simple copy-paste is a strong data augmentation method for instance segmentation.IEEE,CVF Conference on Computer Vision and Pattern Recognition (CVPR),2021:2917～2927.
[14] Chen H,Qi Z,Shi Z.Remote sensing image change detection with transformers.IEEE Transactions on Geoscience and Remote Sensing,2022,60:1～14.
[15] Li Z,Tang C,Liu X,et al.Lightweight remote sensing change detection with progressive feature aggregation and supervised attention.IEEE Transactions on Geoscience and Remote Sensing,2023,61:1～12.
基本信息:
DOI：
中图分类号:TP751;S157
引用信息:
[1]赵刚,颜立军,邱寅凯等.基于MPANet的遥感图像水土扰动异常区识别算法[J].勘察科学技术,2025,No.261(02):17-22.
基金信息:
基于卫星遥感电网工程水土保持监测及评价方法研究(J2023096)