森林资源是地球上最重要的资源之一，在全球生态、环境及社会经济中扮演着重要的角色。近年来，虽然我国森林资源呈现出增长的良好态势，但是我国森林覆盖率、人均森林面积等远低于世界平均水平，森林资源质量不高、人均占有量不足和分布不均的状况仍未得到根本改变。我国仍然处于森林资源短缺，生态系统脆弱的境况，而且我国森林乱砍乱伐乱建现象屡见不鲜，森林资源每年都遭受到了严重的破坏与威胁，通过林地年度变更调查及时摸清森林资源现状并进行科学、合理规划及保护显得尤为重要。 自2017年起，全国范围内全面实施林地年度变更调查工作。由于刚刚起步，不同地区、林业局技术手段不一，虽然已出现基于3S的林地变更调查系统，但遥感（RS）技术应用不够深入。通过利用信息技术、大数据技术等高新技术，建立以遥感（RS）为主的3S集成的林地变更调查系统，能够较为高效地实现林地变更调查，从而服务于全国林地变更调查工作，完成林地“一张图”的更新，保证林地“一张图”的现势性，为森林资源经营管理和生态保护提供决策依据。 本文以全国林地变更调查工作为契机，充分调查研究了当前林地变更流程及相关技术，分析了目前林地变更调查工作和调查系统发展现状及不足。结合林地变更法律法规、政策措施、技术要求，针对目前林地变更存在的问题，建立了以遥感（RS）为主，包括地理信息系统（GIS）和全球定位系统（GPS）的3S集成的林地变更调查系统。调查系统由桌面端年度更新系统和移动端外业数据采集系统两部分共同组成。桌面端年度更新系统采用C、C++和C#等编程语言，结合ArcGIS Engine开发、IDL开发、插件式编程等技术实现，负责完成数据处理与分析等内业工作；移动端外业数据采集系统采用C、C++和Java等编程语言，结合Android开发、移动GIS等技术实现，负责完成林地变更调查数据的现地调查与核实等外业工作。 调查系统集成了3S技术，综合应用了图像处理、基于遥感影像分类后处理的森林变化检测、基于RGB-NDVI图像的森林变化检测、空间分析等技术，实现了从两期遥感影像中快速提取林地变化范围。最后，文本以江西省新余市分宜县为实验区域，对调查系统进行了应用实验验证，基于遥感影像分类后处理的森林变化检测的提取精度达到了81.77%，基于RGB-NDVI图像的森林变化检测的提取精度达到了82.27%。通过实验对比显示，基于RGB-NDVI图像的森林变化检测不需要调查样地数据，且操作简单，精度较高，利于推广和实际应用，可辅助于全国林地年度变更调查工作。

Forest resources are one of the most important resources on the earth and play an important role in global ecology, environment and social economy. In recent years, although China's forest resources have shown a good trend of growth, but China's forest coverage rate, per capita forest area, etc. far below the world average level, the quality of forest resources is not high, the per capita possession of the lack of and uneven distribution of the situation has not yet been Change radically. China is still in a situation of shortage of forest resources and a fragile ecosystem. Moreover, it is not uncommon for China’s forests to be unearthed indiscriminately. Forest resources have been seriously damaged and threatened every year. Scientific, rational planning and protection are particularly important. Since 2017, the annual change investigation of forest land has been fully implemented nationwide. Since the beginning, different regions and forestry bureaus have different technical means. Although the 3S-based forestry change investigation system has emerged, remote sensing (RS) technology is not sufficiently applied. Through the use of high technologies such as information technology and big data technology, a 3S integrated forestry change survey system based on Remote Sensing (RS) will be established, which will enable forestry change investigations to be more efficient, thus serving the nationwide forestry change investigation work and completing forest lands. The update of "a map" ensures that the current situation of "a map" in forestland provides a basis for decision-making on the management and ecological protection of forest resources. This paper took the national forestland change investigation work as an opportunity to fully investigate the current forestland change process and related technologies, and analyzed the current status and shortcomings of forestland change investigation work and survey system development. In light of the changes in laws, regulations, policies, and technical requirements of forest lands, in response to the current problems of forest land changes, 3S integrated forestry changes based on Remote Sensing (RS), including Geographic Information System (GIS) and Global Positioning System (GPS), have been established. Survey system. The survey system consists of a desktop-side annual update system and a mobile-side field data collection system. The desktop-side annual update system uses programming languages such as C, C++, and C#, combined with technologies such as ArcGIS Engine development, IDL development, plug-in programming, etc., and is responsible for completing internal processing such as data processing and analysis; and the mobile terminal field data acquisition system adopts C. Programming languages such as C++, Java, and Java are integrated with technologies such as Android development and mobile GIS. They are responsible for completing fieldwork such as field surveys and verifications of forestry change survey data. The survey system integrates 3S technology, comprehensively applies image processing, forest change detection based on remote sensing image classification post-processing, forest change detection and spatial analysis based on RGB-NDVI images, and realizes rapid extraction of forest land from two remote sensing images. variation range. At last, the text takes the experimental area of Fenyi County, Xinyu City, Jiangxi Province as an experimental area, and the application of the survey system is validated. The extraction accuracy of forest change detection based on remote sensing image classification post-processing reaches 81.77%, and the forest change detection based on RGB-NDVI images. The extraction accuracy reaches 82.27%. The comparison of experiments shows that the detection of forest changes based on RGB-NDVI images does not require investigation of sample data, and the operation is simple, with high precision, which is conducive to promotion and practical application, and can assist in the annual change investigation of forest land throughout the country.