剖面形态自适应的海岸线遥感推算方法

淤泥质海岸冲淤变化大，岸滩剖面形态多样。本文首先根据多时相遥感水边线之间的潮差关系自动判断岸滩剖面形态，进而分别采用不同的函数进行剖面拟合，构建了一种剖面形态自适应的海岸线遥感推算新方法，并在江苏中部淤泥质海岸进行了实证应用。研究表明：下凹形侵蚀岸段、斜坡形平缓岸段和上凸形淤长岸段分别采用三指数衰减函数、线性函数和二阶多项式函数具有良好的剖面拟合效果，利用3条水边线数据拟合所得剖面平均坡度绝对误差分别为0.20‰、–0.17‰和0.13‰，小于剖面实测平均坡度一个数量级。利用5条水边线数据拟合进行海岸线推算时，侵蚀岸段、平缓岸段的海岸线平面位置误差分别为6.5 m和–91.96 m，与平均坡度法相比，误差减小约82.4%。进一步考虑岸滩季节性变化时，使用冬季的水边线数据推算海岸线，对侵蚀岸段和淤长岸段影响不大，但对斜坡形平缓岸段，误差减小约63.65%，因此使用冬季的水边线数据比不区分季节具有更高的海岸线推算精度。

Remote sensing prediction method of coastline based on self-adaptive profile morphology

The muddy coast has a large change in scouring and silting, and the beach profile is diverse. Firstly, according to the tidal range relationship between muti-temporal remote sensing watelines, the shape of the shoreline is automatically judged, and then the different functions are used to fit the profile. A new method of coastline remote sensing prediction based on self-adaptive profile morphology is constructed. The central muddy coast in Jiangsu has been empirically applied. The research shows that the concave-shaped erosion shore section, the slope-shaped gentle bank section and the upper convex-shaped siltation section use a three-exponential decay function, a linear function and a second-order polynomial function respectively to have a good profile fitting effect, using three waterlines. The absolute slope error of the profile obtained by data fitting is 0.20‰, –0.17‰, and 0.13‰, respectively, which is less than an order of magnitude than the measured average slope. When using the five waterlines data fitting to calculate the coastline, the error of the coastline plane position of the erosion shore section and gentle shore section are 6.5 m and –91.96 m, respectively, and the error is reduced by about 82.4% compared with the average slope method. Further consideration of seasonal changes in the beach, using the waterline data of the winter to calculate the coastline, has little effect on the erosion of the shore and the long section of the silt, but for the slope-shaped smooth section, the error is reduced by about 63.65%, so the use of winter waterline data has a higher shoreline projection accuracy than the season without distinction.