基于随机森林算法的濉溪县冬小麦信息提取研究

Research on Winter Wheat Information Extraction in Suixi County Using Random Forest Algorithm

摘要: 基于GF-1遥感影像数据，结合随机森林算法，对安徽省淮北市濉溪县冬小麦种植区域进行了长时间监测研究。通过对研究区2023年10月至2024年6月间15期高分辨率影像进行预处理，主要包括辐射定标、大气校正等操作，生成了濉溪县NDVI和EVI时序影像，分析了濉溪县冬小麦的生长特性。在此基础上，选取关键期影像进行分类，采用随机森林法精准地提取了濉溪县冬小麦种植面积，并与其他分类算法(最大似然法、支持向量机法、神经网络法)进行了精度比较分析。结果表明，随机森林算法在分类精度和稳定性方面具有显著优势，尤其在冬小麦越冬期的提取效果最佳。研究成果可为濉溪县冬小麦种植布局优化和农业监测提供科学依据，验证了随机森林算法在农作物遥感监测中具有很大潜力。

Abstract: Based on GF-1 remote sensing imagery and the Random Forest Algorithm, a long-term monitoring study of winter wheat planting areas in Suixi County, Huaibei City, Anhui Province was carried out. Fifteen high-resolution images of the study area acquired from October 2023 to June 2024 were preprocessed—mainly through radiometric calibration and atmospheric correction—to generate NDVI and EVI time-series images of Suixi County, and the growth characteristics of winter wheat were analyzed. On this basis, images from key growth stages were selected for classification, and the Random Forest method was applied to accurately extract the winter wheat planting area of Suixi County; its accuracy was then compared with that of other classification algorithms (Maximum Likelihood, Support Vector Machine, and Neural Network). The results show that the Random Forest Algorithm has significant advantages in classification accuracy and stability, achieving the best extraction performance particularly during the overwintering stage of winter wheat. These findings can provide a scientific basis for optimizing winter wheat planting layout and agricultural monitoring in Suixi County, demonstrating the great potential of the Random Forest Algorithm in crop monitoring based on remote sensing.