Abstract: With the implementation of integration and optimization of natural protected areas, as well as the progress of boundary surveying and piles setting, the requirements for current status and accuracy of boundary surveying results are increasing. In order to solve the problems in traditional demarcation work, such as not advanced monitoring technology, labor-intensive work, inaccurate results, and unobjective boundaries, etc. This paper proposes a research proposal on the basis of drone tilt photogrammetry, Google map and digital orthophoto, human surveying is liberated while combining the photogrammetry and field surveying, and the boundaries of nature reserve that meet the accuracy requirements would be rapidly identified and extracted. A case study of Beijiang special rare fish provincial nature reserve in Shaoguan is performed, the 1:500 real-scene 3D model is obtained, which could possibly provide clear and reliable background data for the work of "building a special ecological reserve in northern Guangdong with high standards, and do well in the optimization and integration of nature reserves". The results show that the technology is advanced and mature, with a high degree of automation, and can meet the accuracy requirements of boundary surveying. It has broad exploration potential and application prospects, and is easy to be promoted.

Key words: tilt photogrammetry, nature reserve, UAV, demarcation surveying, real scene 3D model