论文部分内容阅读
针对水文预报领域冰层厚度检测过程中对空气与冰层、冰层与冰下水层的分界点难以判断的难题,提出了一种基于介质温度数值分布规律的冰层厚度检测分析方法.新方法在对空气与冰层、冰层与冰下水层的分界点进行分析判断时采用了聚类分析思想,通过采用改进属性归类规则的K-means算法,对采集到的空气、冰与水不同介质的温度数值分布数据进行分类,并对各自属性的温度梯度分布数据集合进行线性拟合,从各拟合曲线交点可以获得空气与冰层、冰层与冰下水层的分界点,进而计算出冰层厚度数值.采用这一方法对2014.12-2015.3内蒙包头黄河河道冰情采集数据进行分析,证明了新方法的可行性.
Aiming at the difficulty in judging the demarcation point between air and ice layer, ice layer and ice layer in the process of ice thickness detection in hydrological forecasting, a new ice thickness detection and analysis method based on the numerical distribution of medium temperature is proposed. Based on the analysis of the demarcation points between air and ice layer, the ice layer and the ice layer, the clustering analysis method is adopted. By using the K-means algorithm which improves the attribute classification rules, the collected air, ice and water are different Medium temperature distribution data were classified and the temperature gradient distribution data set of each attribute was linearly fitted. From the intersection point of each fitting curve, the demarcation point between air and ice layer, ice layer and ice layer under water layer can be obtained, Ice thickness data.Using this method, we analyzed the acquisition data of ice conditions in the Baotou-Huangtang River in Inner Mongolia from December 2014 to December 2015, and proved the feasibility of the new method.