基于机载细分光谱遥感影像的中小河道水体提取  

Extraction of Small and Medium-sized River Water Bodies Based on Airborne Subdivision Spectral Remote Sensing Images

作　　者：崔玉莹 Cui Yuying(Shanghai Shengtu Remote Sensing Engineering Technology Co.,Ltd.,Shanghai,China)

机构地区：[1]上海盛图遥感工程技术有限公司,上海

出　　处：《科学技术创新》2024年第23期101-104,共4页Scientific and Technological Innovation

摘　　要：卫星遥感影像被广泛应用于水质监测领域,在提取大型湖泊河流的水体边界、水体面积方面具有快速准确等优势,受分辨率影响,影像中存在混合像元等现象,在提取细小河流时容易出现误判情况。本文以太浦河支流部分河段为研究对象,分别使用NDWI、NDVI两种指数模型从高分辨率的细分光谱成像设备的高光谱影像中提取中小河道水体信息,分析同一影像两种指数提取结果的不同和同种水体指数不同阈值提取效果的差异性,并通过总体精度、识别水体面积等精度评价指标进行精度分析。结果表明,对于高分辨率的细分光谱遥感影像,阈值0.0008的NDWI水体指数模型的提取效果最好,总体精度达95.6932%,整体效果优于NDVI植被指数。Satellite remote sensing images are widely used in the field of water quality monitoring,which have advantages of fast and accurate extraction of water boundary and water area of large lakes and rivers.Due to the influence of resolution,there are mixed pixels and other phenomena in the images,which are prone to misjudgment in the extraction of small rivers.This paper takes some reaches of Taipu River tributaries as the research object,and uses NDWI and NDVI index models to extract the water body information of small and medium-sized river channels from the hyperspectral images of high-resolution subdivision spectral imaging equipment,and analyzes the difference in the extraction results of two indexes of the same image and the difference in the extraction effects of different thresholds of the same water body index.The accuracy was analyzed by the overall accuracy,the area of the identified water body and other precision evaluation indexes.The results show that for high-resolution subdivision spectral remote sensing images,NDWI water body index model with a threshold value of 0.0008 has the best extraction effect,with an overall accuracy of 95.6932%,and the overall effect is better than NDVI vegetation index.

关 键 词：机载细分光谱 中小河道 NDWI NDVI 水体提取 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]