自党的十八大以来，为加强生态文明建设，我国从2017年开始，要求每年以县为单位进行林地变更调查和林地一张图数据库更新，以便有效监督和查处违法占用林地的事件发生。林地变更调查要求基于互联网+高分遥感、卫星定位和GIS技术，以提高林地变更调查效率、准确性和可靠性。 造成林地变更的主要原因包括：造林、采伐、森林火灾、森林病虫害和林地征占五个方面。林地变更调查的实际作业流程包括：利用两期高分遥感影像人工室内判读林地变化的疑似图斑、派人到现地核实疑似图斑、根据核实结果更新林地一张图数据库三个主要方面。根据全国范围内连续三年林地变更调查实际工作情况，已经暴露出劳动强度大、工作任务紧，可靠性和准确性均有待提高等问题。如何利用互联网+高分遥感、卫星定位和GIS软件开发，提高林地变更调查效率、精度和可靠性，已成为林业科技工作者关注的热点。本文围绕林地变更调查涉及的关键技术问题，完成了以下几个方面的研究与开发工作： （1）根据两期高分遥感影像，研究能够自动提取反应林地动态变化疑似图斑的算法，并进行软件实现。比较算法自动提取的疑似图斑和人工解译的疑似图斑，精度能达到80%左右，可极大减轻室内人工判读的工作量。 （2）提出基于互联网+内外业一体化林地变更调查作业模式，参与网络版的林地变更调查软件平台开发，以省为单位，将林地变更调查所需疑似图斑、两期高分遥感影像和林地规划利用等数据上传至各省的中心数据库服务器，利用网络推送给全省各地市县的用户。通过网络实现任务分发、确保全省调查数据的统一和标准、调查任务的进度管理和调查成果的在线回收。 （3）参与开发移动端调查软件。利用平板电脑，通过用户名和密码登录服务器，下载省监测中心推送的调查所需数据。采用卫星动态导航定位，快速到达疑似图斑现场，进行林地动态变化的核实调查，包括图形和属性数据的采集，拍照取证，轨迹记录、调查数据实时回传和入库等。借助该软件可极大提高野外现场核查的精度和效率，确保调查数据的可靠性和准确性。 利用本文研究和开发成果，于2017年和2018年在四川全省进行了网络化林地变更调查试验，结果表明：利用本文提供的图像检测算法可以减少60%左右的室内人工判读工作量，利用内外业一体化软件平台和数字化作业模式，可以有效提高林地变更调查的效率、精度和可靠性。本文研究和开发成果对我国进行常态化林地变更调查和林地一张图数据库更新具有重要的参考价值。

Since the 18th National Congress of the Communist Party of China, in order to strengthen the construction of ecological civilization, starting from 2017, China requires that forest land change surveys and forest land map databases be updated every county in order to effectively monitor and investigate the occurrence of illegal occupation of forest land. Forest land change survey requirements are based on Internet + high-scoring remote sensing, satellite positioning, and GIS technology to improve the efficiency, accuracy and reliability of forest land change surveys. The main causes of forest land change include: afforestation, logging, forest fires, forest pests, and forest land acquisition. The actual operation process of the forest land change investigation includes three major aspects: the use of two periods of high-score remote sensing images to artificially interpret the suspected spots on the forest land, sending people to the site to verify the suspected spots, and updating the database of a forest land map based on the verification results. According to the survey of the actual work of the forest land change for three consecutive years nationwide, problems such as high labor intensity, tight work tasks, and need to improve reliability and accuracy have been exposed. How to use the Internet + high-scoring remote sensing, satellite positioning, and GIS software development to improve the efficiency, accuracy, and reliability of forest land change investigations has become a hot topic for forestry science and technology workers. This article focuses on the key technical issues involved in forest land change investigations, and completes the following aspects of research and development: (1) Based on the two-phase high-scoring remote sensing images, an algorithm capable of automatically extracting suspected spots that reflect the dynamic changes of forest land is researched and implemented in software. Comparing the suspected speckles automatically extracted by the algorithm and the artificially interpreted suspected speckles, the accuracy can reach about 80%, which can greatly reduce the workload of indoor manual interpretation. (2) Propose the operation mode of forest land change investigation based on the Internet + integration of internal and external industries, participate in the development of an online version of the forest land change investigation software platform, and take the province as the unit to investigate suspected patches and two high-scoring remote sensing images required for forest land change investigation Data such as forest land planning and utilization are placed on the central database server of each province, and they are pushed to users in cities and counties throughout the province by using the network. Realize mission distribution through the network, ensure uniformity and standards of survey data across the province, progress management of survey missions, and online retrieval of survey results. (3) Participate in the development of mobile survey software. Use a tablet computer to log in to the server with a user name and password, and download the survey data pushed by the provincial monitoring center. Adopt satellite dynamic navigation and positioning to quickly reach the spot of suspected spots and conduct verification and investigation of forest land dynamic changes, including the collection of graphics and attribute data, photographs and evidence, trajectory records, survey data real-time return and storage. The software can greatly improve the accuracy and efficiency of field inspections and ensure the reliability and accuracy of survey data. Using the research and development results of this article, networked forest land change survey experiments were conducted in Sichuan Province in 2017 and 2018. The results show that the image detection algorithm provided in this article can reduce the indoor manual interpretation workload by about 60%, using both internal and external It also integrates software platform and digital operation mode, which can effectively improve the efficiency, accuracy and reliability of forest land change investigation. The research and development results of this paper have important reference value for the investigation of normalized forest land change and the update of a map database of forest land in China.