%0 Journal Article %T 基于多源遥感数据融合的农村闲置宅基地目标识别研究
Research on Target Recognition of Rural Idle Homestead Land Based on Multi Source Remote Sensing Data Fusion %A 郑荣宝 %A 陈美招 %A 陈晶 %A 唐晓莲 %J Geographical Science Research %P 464-472 %@ 2168-5770 %D 2025 %I Hans Publishing %R 10.12677/gser.2025.142046 %X 闲置宅基地的目标识别是土地科学界难以解决的技术问题，我国到底有多少农村闲置宅基地一直缺乏准确数据。本文基于多源遥感数据融合的数据处理方法，通过高分夜光遥感数据与无人机航飞数据的耦合，实现农村闲置宅基地的目标精确识别。研究结果表明：① 结合高分辨率遥感影像和夜间灯光数据，基于支持向量机分类方法中，Kappa系数分别为82.09%和0.8045，基本达到目标识别精度要求。② 惠东县的农村闲置宅基地总体比例较高，超过宅基地总面积的三分之一，实验证明了基于高分夜光遥感影像与高分航飞影像的数据融合进行农村闲置宅基地识别的研究是可行的。③ 需要创新农村闲置宅基地管理制度与方法，实现农村宅基地的盘活利用，以提高农村建设用地利用效率。本文提供了一种高效快速的农村闲置宅基地的目标识别方法，可为农村低效闲置用地改造利用提供技术支撑。
It is a very difficult technical problem which is to solve target recognition of idle homesteads in land science. Not everyone knows the accurate data on how many idle rural homestead lands exist in China. Based on a data processing method of multi-source remote sensing data fusion, this article gets the accurate identification of idle rural homestead lands through the coupling of high-resolution night light remote sensing data and unmanned aerial vehicle (UAV) flight data. The research results show that: ① Combining high-resolution remote sensing images and nighttime light data, the Kappa coefficients in the support vector machine classification method are 82.09% and 0.8045, respectively, which basically meet the requirements of target recognition accuracy. ② The overall proportion of idle rural homestead lands in Huidong County is relatively high, exceeding one-third of the total area of homestead lands, and it has been proven that the research on identifying idle rural homestead lands based on the fusion of high-resolution night light remote sensing images and high-resolution aerial images is feasible. ③ It is necessary to innovate the management system and methods of idle rural homestead lands, realize the revitalization and utilization of rural homestead lands, and improve the efficiency of rural construction land utilization. The research results of this article provide an efficient and fast target recognition method for idle rural homestead lands, which can provide technical support for the transformation of inefficient idle rural land. %K 图像融合， %K 闲置宅基地， %K 目标识别
Image Fusion %K Idle Homestead Land %K Target Recognition %U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=113349