近岸海域水色异常遥感监测方法研究

Study on Remote Sensing Monitoring Method of Water Color Anomaly in Coastal Waters

水体颜色是判断水体质量的重要依据,利用遥感技术检测水色异常是进行水质评估和水体污染监测的重要技术手法。基于盐城市废黄河入海口处的GF1-WFV影像数据,利用遥感影像自身光谱特征来构建水色异常判断函数,对研究区域的水色异常进行快速提取和定位,并比较分析了基于单景影像特征和多景影像特征设置判断阈值的提取结果。研究结果表明:该方法完全基于遥感影像光谱特征,可有效提取出未知类型的水色异常信息;与基于单景影像设置的判断阈值相比,基于双景影像设置的判断阈值更合理、更适用,提取速度更快,提取范围更精确,Kappa系数为0.80;另外,采用合适窗口大小进行均值滤波处理能有效防止提取结果细碎和斑块化。

Water color is an important basis for judging water quality.Remote sensing for water color anomaly detection is an important technique for water quality assessment and water pollution monitoring.Based on the GF1-WFV image data at the estuary of the waste Yellow River in Yancheng City,the judgment function of water color anomaly is constructed by using the spectral characteristics of the remote sensing image itself,and the water color anomaly in the research area is quickly extracted and positioned.The extraction results of the judgment threshold based on single-view image features and multi-view image features are compared and analyzed.The results show that this method is completely based on spectral characteristics of remote sensing images and can effectively extract unknown water color anomalies.Compared with the judgment threshold based on a single-view image,the judgment threshold based on a double-view image is more reasonable and applicable,the extraction speed is faster,the extraction range is more accurate,and the Kappa coefficient is 0.80;In addition,averaging filtering with appropriate window size can effectively prevent fragmentation and patching of extraction results.