基于多源高分遥感影像的围海养殖动态监测

Dynamic Monitoring of Coastal Aquaculture Ponds with Multi-source High Spatial Resolution Imagery

围海养殖是一种重要的海水养殖方式,在产生巨大经济效益的同时,也为我国近海生态环境、海上交通等方面带来负面影响#动态监测围海养殖区,对于高效开发利用沿海渔业资源、保护海洋生态环境、以及掌握海洋灾害承灾体信息意义重大。文章以湛江市北莉岛为研究区,基于2019年6月22日高分一号卫星影像和10月29日资源三号02星影像,使用支持向量机自动分类算法对研究区内正在养殖的围海养殖区范围进行监测与分析。2019年6月和10月高分辨率卫星数据的围海养殖区提取正确识别率分别为94.47%和92.98%,面积分别为266.67 ha和271.70 ha。总体而言,研究区两个时相正在养殖中的围海养殖区范围变化很小。但本研究能较好地区分和监测正在养殖的池塘和干涸池塘,为海域使用精细化管理、渔业估产等提供基础数据支撑。

Coastal aquaculture is an important way of mariculture,which not only produces huge economic benefits,but also brings negative impacts on offshore ecological environment and maritime traffic.Dynamic monitoring of coastal aquaculture ponds is of great significance for the efficient development and utilization of coastal fishery resources,protection of the marine ecological environment as well as information collection of marine hazard-bearing bodies.This article chooses Beili Island in Zhanjiang city as the research area,combines the Gaofen-1 satellite image acquired on June 22,2019 and the Ziyuan-302 satellite image acquired on October 29,and uses the support vector machine classification algorithm to analyze and monitor the coastal aquaculture ponds in use in the study area.Based on Gaofen-1 and Ziyuan-302 satellite images,the extraction accuracies of the coastal aquaculture areas are 98.73%and 98.32%respectively;the areas of the detected coastal aquaculture ponds in use are 266.67 ha and 271.70 ha respectively.In general,there is little area change of the aquaculture ponds in use in the study area during the four months of 2019.However,this study can distinguish and monitor the pond areas in use with water and the dry aquaculture areas,providing basic data support for the refined management on sea area use and fishery production estimates.