基于国产高分卫星GF-1影像的农膜遥感信息提取技术研究

Study on Extraction of Plasticulture Basing on GF-1 of High Resolution Image of China

随着农膜的大量使用,农膜对土壤的污染已被广泛关注。准确掌握农膜空间分布信息和使用量,可以为农膜环境污染治理提供重要技术支撑,也为政府调节市场和发放设施农业补贴提供重要依据。传统的农膜使用量统计方法费时费力,没有空间定位信息。基于遥感技术农膜自动化提取则具有快速、动态、空间的优势。本研究以山东省寿光市为例,采用国产GF-1卫星遥感影像,利用支持向量机(SVM)分类技术,进行农膜遥感信息提取。经检验,Kappa系数达96%,达到解译精度要求。寿光市覆膜农田面积397km^2,占全市农田面积的33. 6%,农膜使用量3. 34万t,农膜主要分布在寿光市中部和南部。结果表明,利用GF-1卫星遥感影像能较好地提取农膜信息。

Along with extensive use of plasticulture,contamination of the soil by plasticulture has been widely concerned.Extracting accurately the spatial distribution and the amount of plasticulture,which can provide important technical support for the governance of plasticulture pollution,which can provide an important basis for the administration of market regulation and subsidies distributing of facility agricultural.Traditional statistical methods of plasticulture use are time-consuming.Automation extraction of plasticulture basing on remote sensing technology are fast,dynamic,along well with the spatial distribution of plasticulture.The plasticulture of Shouguang City of Shandong province was extracted by using support vector machine classification method with GF-1image of China.By inspection,the Kappa coefficient was96%,which can achieve the accuracy of interpretation requirement.The area of farmland mulched with plasticulture in Shouguang city was397km^2,accounts for33.6%of city’s farmland area.The usage amount of plasticulture was33400ton.The distribution of plasticulture was in the middle and south of Shouguang City.The results showed that it could extract preferably plasticulture with GF-1satellite remote sensing image.