关键词: BP神经网络, 主成分分析, 高光谱, 植被指数, 叶绿素含量, 红边参数

Abstract: Land ecological security, the core of sustainable utilization of land resource, is relevant to human well-being. The study of land ecological security is of great significance to reveal the characteristics of land use security and promote the coordinated regional development of population, resources and environment. Using the “PressureState-Response (PSR)” model, we constructed an evaluation index system of land ecological security in Fenhe River Basin. The index weight was determined by AHP and entropy method, while the matter-element model was used to evaluate land ecological security. In 2006, the land ecological security of Fenhe River Basin was at the state of “transforming to safe status”. In 2012 and 2017, it was at the state of “transforming to moderately safe status”, with overall situation being stable. From 2006 to 2017, landscape fragmentation, per capita water resource and river network density were the main factors restricting land ecological security. In 2017, 15% of the counties (cities/districts) were at the “safe” state, with most of which being concentrated in the upstream area. Land ecological security in the upstream area was generally higher than that in the middle and downstream areas.

Key words: hyperspectral, red edge parameters, vegetation index, principal component analysis, BP neural network., chlorophyll content