论文部分内容阅读
以江苏大丰高涂围海养殖区域为研究区,基于ZY-3号高分辨率卫星遥感数据,采用面向对象方法开展了围海养殖用海信息分类提取与动态监测研究。通过多尺度分割获取研究区地物斑块,根据其形状、光谱、纹理等特征信息的差异性,建立分类规则集,并结合目视解译完善提取规则,实现了围海养殖用海中淡水养殖区、海水养殖区和干涸养殖区的分类提取,得到精度达93%的分类结果。总结出养殖区分类提取通用的特征指标集。最后基于目标识别,对多年份围海养殖信息进行动态监测,更全面、高效地了解围海养殖发展趋势,为海洋遥感监测领域拓展了更广阔的应用前景。
Based on the ZY-3 high-resolution satellite remote sensing data, the Dafeng high-rise marine culture area in Jiangsu Province was used as the research area to carry out the information classification extraction and dynamic monitoring of the sea reclamation using the object-oriented method. According to the difference of the characteristic information such as shape, spectrum and texture, the classification rule set was established by multi-scale segmentation, and the perfect extraction rules were combined with visual interpretation to realize the marine freshwater aquaculture District, mariculture zone and dry farming area of the classification extraction, accuracy of 93% of the classification results. Summed up the general characteristics of the aquaculture zone extraction feature set. Finally, based on the target identification, dynamic monitoring of multi-year marine rearing information is carried out to understand the development trend of marine rearing in a more comprehensive and efficient manner and broaden the broader application prospects for the monitoring of marine remote sensing.