UAV Hyperspectral Remote Sensing System Integration And Grassland Feature Recognition

In the national standard,the area change of vegetation and bare soil is regarded as an important index to evaluate the grassland degradation grade.At present,the means of grassland degradation monitoring and research are manual survey and satellite remote sensing.Manual survey is time-consuming and laborious,and cannot carry out real-time monitoring and quantitative research.Due to the low resolution and limited accuracy of satellite remote sensing,many grassland degradation indicators cannot be applied.Therefore,there is an urgent need for a flexible and convenient grassland degradation monitoring and research method that can perform real-time monitoring and quantitative research.In this study,the key components of the UAV hyperspectral remote sensing system were designed,and the UAV hyperspectral remote sensing system was constructed.The system was used to collect data in the grassland field,and the low-altitude hyperspectral identification and classification of grassland bare soil and vegetation were realized.The pan-tilt is an important component of the UAV hyperspectral remote sensing system.It connects the UAV and the hyperspectral spectrometer,which directly affects the data acquisition quality of the hyperspectral spectrometer.Therefore,its design is the key content of the UAV hyperspectral remote sensing system integration.Due to the strong air convection over the grassland,the stability of the common cloud platform is insufficient,and it is easy to cause image distortion,which puts forward higher requirements for the design of the cloud platform.In this study,two schemes of three-axis stabilized platform were designed.Based on various factors,the design scheme of bistable rabbit cage structure was determined,and its three-dimensional model was constructed by Solid Works software.Then through animation analysis and Motion analysis,the angular velocity and force of the parts are graphically analyzed.The results show that the requirements can be met,and there is no interference in the space motion stroke.In this study,the original grassland in Hunshandake Sandy Land of Zhenglan Banner,Inner Mongolia was taken as the research area.Based on the designed cloud platform,a UAV hyperspectral remote sensing system was constructed to collect hyperspectral data of grassland features under natural light.The hyperspectral data were input into the software ENVI5.3 after reflectance correction.The spectral characteristics of vegetation in the image were extracted by spectral differential method and envelope removal method respectively.According to the different parametric characteristics,the vegetation in two typical test areas was successfully distinguished.Then,four kinds of vegetation index threshold methods were used to analyze and process hyperspectral data,and the pixel value distribution and segmentation threshold of the three objects under the vegetation index gray image were obtained.After the accuracy verification of the Kappa coefficient confusion matrix,it was concluded that the MSAVI threshold method had the highest classification accuracy of vegetation and bare soil.This study can provide technical support and analysis methods for estimating the vegetation coverage of desertification grassland and judging the process of grassland degradation succession.