基于CBERS-04卫星影像的珊瑚礁底质分类能力分析

Analysis of Classification Ability of Coral Reefs Based on CBERS-04 Satellite Imagery

珊瑚礁对于海洋生态环境研究具有重要意义,通过分析珊瑚礁的分布以及生长状况可以对海洋生态环境进行评估。本文以南海珊瑚岛近岸海域为研究区域,选取中巴地球资源四号(CBERS-04)卫星影像数据,利用支持向量机(SVM)分类方法对近岸珊瑚礁底质进行分类,分类结果进行矢量化输出,对珊瑚礁数量和面积进行统计,与Worldview-2卫星多光谱影像数据分类矢量结果进行对比分析。实验结果表明,利用中巴地球资源卫星数据对于珊瑚礁分类是可行的,其中利用20 m空间分辨率多光谱数据与5 m空间分辨率全色波段数据融合后的影像数据,对于珊瑚礁信息提取效果最好。但由于空间分辨率及波段光谱设置的限制,其对于浅水区小面积珊瑚礁和深水区珊瑚礁提取能力较弱。

Coral reefs are of great significance for the study of marine ecological environment.The marine ecological environment can be evaluated by analyzing the distribution and growth of coral reefs.This paper takes the coastal waters of the South China Sea coral island as the study area,and selects the CBERS-04 satellite image data,using the support vector machine(SVM)classification method to classify the nearshore coral reefs,output classification vector results,statistics on the number and area of coral reefs,and compare the results of Worldview-2 satellite multispectral image data classification vector results.The experimental results show that the use of China-Brazil Earth Resources Satellite data is feasible for coral reef classification.The image data obtained by combining 20 m spatial resolution multi-spectral data with 5 m spatial resolution panchromatic data is the best for coral reef information extraction.However,due to the limitation of spatial resolution and spectral setting of the band,it has weak ability to extract coral reefs in deep water areas and coral reefs in deep water areas.