基于多光谱图像融合的区域光伏发电容量预测  

REGIONAL PHOTOVOLTAIC POWER GENERATION CAPACITY PREDICTION BASED ON MULTISPECTRAL IMAGE FUSION

作　　者：马晓磊 张彦军 汪凯威[1] 孙林华 李永光 Ma Xiaolei;Zhang Yanjun;Wang Kaiwei;Sun Linhua;Li Yongguang(State Grid Xinjiang Electric Power Co.,Ltd.,Urumqi 830063,China;College of Electrical Engineering,Sichuan University,Chengdu 610065,China;Guodian Nanrui Nanjing Control System Co.,Ltd.,Nanjing 211102,China)

机构地区：[1]国网新疆电力有限公司,乌鲁木齐830063 [2]四川大学电气工程学院,成都610065 [3]国电南瑞南京控制系统有限公司,南京211102

出　　处：《太阳能学报》2024年第11期267-271,共5页Acta Energiae Solaris Sinica

摘　　要：为了提高电力系统运行的安全性,提出一种基于多光谱图像融合的区域光伏发电容量预测方法。确定区域光伏发电管辖范围,采集多光谱卫星遥感图像,对图像进行IHS变换和Curvelet变换,通过多光谱图像融合技术融合标准差,分析光伏电站时空特征,集合历史数据,动态预测区域光伏发电容量。实验结果表明:该方法的预测均方误差为0.526、互信息以及结构相似性均在0.9以上,可以清晰呈现光伏电站时空多光谱图像,区域光伏发电容量真实值与预测值较为拟合,可以准确预测区域光伏发电容量的变化情况。In order to improve the safety of power system operation,a region photovoltaic(PV)power generation capacity prediction method based on multispectral image fusion is proposed.The method defines the jurisdiction of regional PV power generation,acquires multispectral satellite remote sensing images,applies IHS transformation and Curvelet transform to the images,and fuses the standard deviation of multispectral images through image fusion technology.By analyzing the spatio-temporal characteristics of the PV power station and combining historical data,the method dynamically predicts the regional PV power generation capacity.Experimental results show that the mean square error of the method is 0.526,and the mutual information and structural similarity are both above 0.9.The method can clearly present the spatio-temporal multispectral images of the PV power station,and the real and predicted values of regional PV power generation capacity are well fitted,which can accurately predict the changes in regional PV power generation capacity.

关 键 词：多光谱 图像融合 CURVELET变换 区域光伏发电 容量预测 

分 类 号：TM615[电气工程—电力系统及自动化]