【摘要】 作为核心遥感卫星产品，土地覆盖产品通过表征地球表面状态与变化，为全球环境和区域可持续发展提供关键数据支撑。然而，受影像数据源异质性、分类技术多样性和生产环境差异性等多重因素影响，现有产品在品种规格、空间分辨率、分类精度、覆盖规模以及更新时效性等方面存在显著差异，导致其真实性检验面临严峻挑战。文章面向“百星一用”航天遥感应用模式，提出一种基于分类控方的土地覆盖产品真实性检验方法。从对检验样本在数量、质量和分布三方面的要求出发，构建分布可控、数量可控、质量可控的可信分类控方，优化检验流程，实现对不同卫星数据所生产的土地覆盖产品开展真值检验，为大规模土地覆盖产品的真实性检验提供了一种新的途径。更多还原

【Abstract】 Land cover products are the main products among remote sensing satellite products, presenting the state and changes of the earth’s surface and serving global environment and regional sustainable development.However, affected by factors such as different image data sources, classification techniques and production environments, the product accuracy shows great differences in practical applications, which are reflected in product varieties, specifications, quality, scale and timeliness, resulting in huge challenges in the authenticity verification of land cover products in practical applications. In this paper, aiming at the aerospace remote sensing application mode of "standarization for satellites design and it’s optimal applications", a land cover product verification method based on ground control DS of classification is proposed. Starting from the requirements for verification samples in terms of quantity, quality, and distribution, it constructs credible classified control parties with controllable distribution, quantity, and quality, optimizes the verification process, and realizes truth-value verification for land cover products produced from different satellite data. By optimizing the verification process,a new approach is provided for the authenticity verification of large-scale land cover products.更多还原