【摘要】 【目的 】耕地复种是重要的农业生产方式，也是衡量耕地资源集约化利用水平与可持续利用能力的关键指标。动态监测其空间分布格局、精准把握其变化规律，是优化农业种植模式、制定农业发展政策、保障国家粮食安全的重要基础。【方法 】文章采用文献综述的方法，系统梳理近年来国内外复种遥感监测的研究方法与典型案例，深入剖析现有方法存在的问题，展望复种遥感监测研究的发展方向，以期为该领域的后续研究提供有益参考。【结果 】现有复种遥感监测方法可分为基于高时间分辨率数据的监测方法、基于特征时相的复种提取方法、多源数据融合的复种提取3类。随着数据源的日益丰富以及计算机技术的快速发展，复种遥感监测方法与技术也取得了长足进步，但现有方法仍面临数据预处理的基础性问题、耕地种植模式难以精准刻画、遥感监测时段与实际生育期不吻合、复种遥感监测缺乏地块级信息集成等问题，且监测流程多依赖于分步处理，易导致误差传递与累积。复种遥感监测的数据源和算法各个环节的不确定性尚未被系统认知与量化，地面验证数据共享亦存在明显短板。【结论 】未来耕地复种遥感监测需依赖多源异构时间序列数据的协同融合，以突破多云雨地区的数据源限制。研究应重点围绕以下几方面展开：针对现有分步流程的误差累积问题，建立复种遥感监测的误差溯源体系与不确定性分析方法；发展“端到端”的复种遥感监测一体化智能监测框架，降低误差传播与累积影响；驱动产品服务化与轻量化应用转型，弥补数据产品与农业生产管理实际需求的鸿沟；构建物候与复种状况智能观测体系及全球共享网络。通过遥感、农业和计算机等多学科交叉与国际协作，进一步推动耕地复种遥感监测技术进步，服务农业可持续发展。更多还原

【Abstract】 [Purpose] Multiple cropping is an important agricultural practice and a key indicator for assessing cultivated land use intensity and sustainable utilization capability. Dynamically monitoring its spatial distribution and accurately understanding its temporal-spatial dynamics are essential for optimizing agricultural cropping systems, formulating effective agricultural policies, and ensuring national food security. [Method] This paper employed a literature review approach to systematically summarize research methods and typical cases in remote sensing-based multiple cropping monitoring,both domestically and internationally,in recent years. It thoroughly analyzed the shortcomings of current methods and future research directions. The objective was to provide useful references for subsequent studies in this field. [Result] Existing remote sensing-based multiple cropping monitoring methods were categorized into three types: methods based on high-temporal-resolution time-series data, extraction methods based on key phenological phases,and approaches utilizing multi-source data fusion. With the increasing availability of remote sensing data and rapid advances in computer science technologies, significant progress was made in remote sensing-based methods for multiple cropping monitoring. However, these methods continued to encounter challenges, such as fundamental issues in data preprocessing, difficulties in accurately characterizing cropping patterns including omission of short crop cycles, uncertainty arising from fallow land and abandoned cropland,misalignment between remote sensing monitoring periods and actual crop growth cycles,and a lack of plot-level information integration. Moreover,monitoring processes often relied on multi-step procedures,which easily led to error propagation and accumulation. Furthermore,the uncertainties associated with the various data processing stages and algorithms for multiple cropping monitoring were not systematically recognized and quantified,and there were significant shortcomings in the sharing of ground validation data. [Conclusion] Effective future monitoring of multiple cropping in cloudy regions will hinge on the integration of multisource time-series data. To mitigate the error propagation and accumulation during cropping intensity monitoring,key research priorities are suggested concentrating on the establishment of systems for error traceability,and the systematic uncertainty analysis and the development of end-to-end integrated intelligent monitoring frameworks for cropping patterns. Additionally,there is a necessity to construct intelligent observation systems and global sharing networks for phenology and cropping status. The advancement of remote sensing-based multiple cropping monitoring,driven by interdisciplinary collaboration among remote sensing,agriculture,and computer science, as well as international cooperation, will ultimately support sustainable agricultural development.更多还原