基于无人机低空遥感的高海拔自然保护区垃圾散布区自动监测研究

随着珠峰绒布寺以上核心区、青海宝玉则国家公园等保护区因生态保护升级而被关停，高海拔自然保护区“陋习”垃圾因无法得到及时处理而造成的水源、植被、土壤、冰川等生态环境破坏已成为了旅游产业可持续性发展的重大问题。针对这一问题，基于无人机低空遥感的垃圾散布区自动监测成为了一项迫切需要解决的关键技术。但是，由于①遥感样本缺乏导致端对端的深度学习技术难以启动；②大量多时相多视角低空遥感图像需要稳定高效的配准；以及③垃圾自身的特点等问题给该技术的研究带来了巨大挑战。本课题拟通过规划合理的无人机航拍，构建稳定高效的图像配准方法，构建有效的垃圾散布候选区自动提取、识别、定位方法，来解决高海拔自然保护区垃圾散布区自动监测的相关技术问题。通过本课题的研究，有望建立一种新型的基于无人机的垃圾散布区自动监测手段，为生态环境保护与旅游产业的可持续发展提供技术支撑，相关研究成果具有重要的理论意义与实际推广应用价值。

With the core area of mount Everest rongbu temple，qinghai baoyu national park and other nature reserves being shut down due to ecological protection upgrades, the main reason is widely considered to be the ecological damages in water source, vegetation, soil, glacier and etc caused by the garbage failed to be cleaned up in time. Such ecological damages have become a major issue in the sustainable development of the tourism industry. For this issue, automatic monitoring of garbage scattered areas using low altitude remote sensing of unmanned aerial vehicle (UAV) has become a key solution that needs to be provided urgently. However, ①the lack of remote sensing training samples makes end-to-end deep learning methods hard to use; ② a large number of multi-temporal and multi-viewpoint low-altitude remote sensing images require a stable and efficient registration; and ③ the characteristics of garbage, have brought great challenges to this research. In this proposal, we try to plan a reasonable UAV aerial photography, build a stable and efficient image registration, and build efficient automatic detecting, recognizing and locating methods of garbage scattered areas for solving problems of automatic monitoring garbage scattered areas in high altitude nature reserves. Through this study, it is expected to establish a new approach for monitoring garbage scattered areas in nature reserves, which may provide a new technical support for the ecological environment protection and the sustainable development of tourism industry, and relevant achievements in this research will have the important theoretical significance and the practical application value.