| The Inner Mongolia section of the Yellow River is located at the northernmost end of the Yellow River Basin.The flood season of this section is from mid-to-late November to mid-March of the following year.The flood season lasts for 3 to 4 months and has the remarkable characteristics of closing the river from bottom to top,opening the river from top to bottom,and gradually closing the river in stages.Affected by factors such as global warming,riverbed evolution,and human activities,the disaster-pregnancy environment in the river reaches is complex,with various disaster-causing factors,strong suddenness,and strong chain occurrence.In the past,scholars’ analysis of the Yellow River flow was often limited to the one-point observation of hydrological stations,and the research contents were mostly the closing time of each river section,the length of closing the river,the thickness of river ice,the increment of water storage in the trough,the local river ice,and ice type distribution and other characteristics.However,few scholars pay attention to extracting the basic data of Ling flood based on remote sensing digital image processing and geographic information system technology.This paper takes the Inner Mongolia section of the Yellow River as the research area,and based on the multi-source time series remote sensing data during the flood season,extracts the distribution,type and closed river information of the river ice during the flood season,and combines high-definition historical remote sensing images,the Yellow River Network releases Ling flood data and field surveys.The optimal extraction thresholds of river ice pixels,river ice types and closing time of rivers in the flood season in the study area were obtained.The research results are as follows:(1)Based on the experimental comparison of different remote sensing indexes,it is concluded that the optimal extraction threshold of river ice pixels in each sub section of the Inner Mongolia section of the Yellow River by NDUWI is between 0.16 and 0.24,and the overall classification accuracy and kappa coefficient are greater than 90.00% and 0.90,which improves the accuracy of river ice interpretation;(2)Based on the dynamic threshold method,it is concluded that the optimal threshold of FUS and BUE in the study river section is between 0.05 sa and 0.20 SA,the extraction deviation is within ± 3 days,and the deviation of FUE and BUS extracted based on slope method is within ± 5 days;(3)Based on the curve slope method and the dynamic threshold method,the optimal threshold for the closed and open river information in the river channel is formulated,and the extraction deviation within the optimal threshold is within 5 days;(4)Based on the polarization characteristic parameters of radar images in recent three years,using Markov distance and minimum distance classifier,the best discrimination threshold of rough ice and flat skating in shisifenzi reach of Tuoketuo county is-15.28 db,and the distribution map of river ice types is drawn;(5)The correlation coefficient between air temperature and the first ice day was 0.81,and the correlation coefficient between flow and water level and the beginning and complete melting days was-0.64;The research results can provide some reference for the overall ice monitoring,field basic data collection,site selection of riverbank embankment in disaster prone reach,ice blasting and field survey location in Inner Mongolia section of the Yellow River. |
PDF Full Text Request