北京测绘2025，Vol.39Issue(8)：1205-1210,6.DOI:10.19580/j.cnki.1007-3000.2025.08.020
基于多特征耦合的广东省垦造水田种植监测
Monitoring of rice paddy cultivation in Guangdong province based on multi-feature coupling
摘要
Abstract
Due to the significant impact of cloudy and rainy weather in southern regions,optical imagery quality is unstable,and the accuracy of rice planting extraction based on a single feature is limited.This paper proposed a technique for identify-ing rice cultivation conditions by coupling time-series synthetic aperture radar(SAR)data with phenological and variance features.First,Sentinel-1 satellite images from March to October 2023 were selected.Using a curve similarity metric,the similarity between each pixel on the Sentinel-1A image and rice sample curves was compared.Then,rice phenology and time-series variance feature parameters were extracted.Finally,these features were coupled and supervised classification was performed using decision trees to extract rice cultivation information.The results show that the multi-feature coupling method enriches the rice recognition information and provides more reliable monitoring results for the reclamation of rice pad-dies.This method outperforms the single curve similarity metric,improving overall accuracy by approximately 6%.The rice cultivation identification method based on time-series radar data and multi-feature coupling improves the extraction accuracy of rice cultivation and has significant implications for the monitoring of rice paddy reclamation.关键词
水稻/哨兵一号卫星/光谱相似值/物候特征/方差特征Key words
rice/Sentinel-1 satellite/spectral similarity value/phenological features/variance features分类
农业科技引用本文复制引用
叶露,吴成明..基于多特征耦合的广东省垦造水田种植监测[J].北京测绘,2025,39(8):1205-1210,6.基金项目
广东省省级科技计划(2021B1111610001) （2021B1111610001）
