Analysis of Unmanned Aerial Vehicle (UAV) hyperspectral remote sensing monitoring key technology in coastal wetland

Abstract

The coastal wetland, a transitional zone between terrestrial ecosystems and marine ecosystems, is the type of great value to ecosystem services. For the recent 3 decades, area of the coastal wetland is decreasing and the ecological function is gradually degraded with the rapid development of economy, which restricts the sustainable development of economy and society in the coastal areas of China in turn. It is a major demand of the national reality to carry out the monitoring of coastal wetlands, to master the distribution and dynamic change. UAV, namely unmanned aerial vehicle, is a new platform for remote sensing. Compared with the traditional satellite and manned aerial remote sensing, it has the advantage of flexible implementation, no cloud cover, strong initiative and low cost. Image-spectrum merging is one character of high spectral remote sensing. At the same time of imaging, the spectral curve of each pixel is obtained, which is suitable for quantitative remote sensing, fine classification and target detection. Aimed at the frontier and hotspot of remote sensing monitoring technology, and faced the demand of the coastal wetland monitoring, this paper used UAV and the new remote sensor of high spectral imaging instrument to carry out the analysis of the key technologies of monitoring coastal wetlands by UAV on the basis of the current situation in overseas and domestic and the analysis of developing trend. According to the characteristic of airborne hyperspectral data on UAV, that is three high and one many, the key technology research that should develop are promoted as follows: 1) the atmosphere correction of the UAV hyperspectral in coastal wetlands under the circumstance of complex underlying surface and variable geometry, 2) the best observation scale and scale transformation method of the UAV platform while monitoring the coastal wetland features, 3) the classification and detection method of typical features with high precision from multi scale hyperspectral images based on time sequence. The research results of this paper will help to break the traditional concept of remote sensing monitoring coastal wetlands by satellite and manned aerial vehicle, lead the trend of this monitoring technology, and put forward a new technical proposal for grasping the distribution of the coastal wetland and the changing trend and carrying out the protection and management of the coastal wetland.