基于厘米级无人机影像的水土保持措施精准识别

基于厘米级高分辨率无人机影像,应用面向对象方法(Object-Based Image Analysis, OBIA)对吉林省伊通县椽子沟流域的横坡改垄、地埂植物带、生态恢复乔木林、生态恢复草地等水土保持措施进行自动精准识别。应用超绿指数(Excess Green Index,ExG)、超红指数(Excess Red Index,ExR)、归一化差异指数(Normalized Difference Index,NDI)等光谱指数,形状的主方向、形状指数等形状特征,均值(Mean)、方差(Variance)、对比度(Contrast)等纹理特征进行措施的特征提取。结果表明:研究区水土保持措施识别的总体精度可达91.24%,Kappa系数为0.87;对垄台、垄沟等线性水土保持措施总体精度可达72.33%,Kappa系数为0.63。基于厘米级无人机影像,应用面向对象方法基本可实现对黑土区水土保持措施的精准识别,也可对垄台垄沟等线性措施进行自动识别,研究结果可为水土保持措施实施范围及完好程度的动态监测提供参考依据。

Accurate Identification of Soil and Water Conservation Measures Based on Centimeter-resolution UAV Images

Based on the centimeter-resolution Unmanned Aerial Vehicles (UAV) image, the Object-Based Image Analysis (OBIA) was applied to identify the soil and water conservation (SWC) measures, including contour ridges, ridge plants, ecological restoration forest, ecological restoration grassland, in the Chuanzigou catchment of Yitong district, Jilin province. Spectral indices including Excess Green Index (ExG), Excess Red Index (ExR), Normalized Difference Index (NDI), shape features including principal direction and shape index, and texture features including mean, variance and contrast were used to extract the features of SWC measures. Finally, the overall accuracy and Kappa coefficient of the identification of SWC measures in the study area was 91.24% and 0.87, respectively. The overall accuracy of the identification of linear SWC measures such as ridges and ridge-furrows was 72.33% with a Kappa coefficient of 0.63. The research showed that the centimeter-resolution UAV image, combined with the OBIA methods had the ability to accurately identify SWC measures in black soil area. They can also identify linear measures such as ridges and ridge-furrows automatically. The research can provide a reference basis for the dynamic monitoring of the implementation scope and integrity of SWC measures.