The prediction of the future position of the shoreline is of great importance in planning studies in coastal areas, in making effective decisions in coastal management, and in determining changes occurring on the coast. In this study, coastal change analyses were performed in two different study areas (the Gulf of Izmit and the Göksu Delta) by using satellite images of different dates, and the accuracy of the Kalman filter-based future shoreline prediction method was determined by statistical methods. In this context, by using the 19-period Landsat satellite image belonging to different dates between 1975 and 2019 for the Gulf of Izmit and the 10-period Landsat satellite image belonging to different dates between 1984 and 2018 for the Göksu Delta, shorelines were extracted automatically, and coastal changes were analyzed at a 95% CI by the statistical methods of end point rate, linear regression rate, and weighted linear regression rate (WLR). Afterward, the shorelines extracted automatically on the determined dates were compared with 10-year and 20-year predicted shorelines by the Kalman filter-based prediction method, and their accuracy was statistically analyzed. As a result, the 10-year predicted shorelines by the WLR method were found to provide the highest accuracy.

中文翻译：

---

基于卡尔曼滤波器的未来海岸线预测方法的评估

在沿海地区进行规划研究，在沿海管理中做出有效决策以及确定沿海地区发生的变化时，预测海岸线的未来位置非常重要。在这项研究中，通过使用不同日期的卫星图像，在两个不同的研究区域（伊兹密特湾和戈克苏三角洲）进行了海岸变化分析，并通过统计方法确定了基于卡尔曼滤波的未来海岸线预测方法的准确性。在这种情况下，通过使用伊兹密特湾1975年至2019年之间不同日期的19个周期Landsat卫星图像和Göksu三角洲1984年至2018年之间不同日期的10个周期的Landsat卫星图像，提取了海岸线自动地 通过终点率，线性回归率和加权线性回归率（WLR）的统计方法，以95％CI对沿海和沿海地区的变化进行了分析。然后，通过基于卡尔曼滤波的预测方法，将在确定日期自动提取的海岸线与10年和20年的预测海岸线进行比较，并对其准确性进行统计分析。结果，发现通过WLR方法预测的10年海岸线具有最高的准确性。