基于特征算法的塔式光热电站镜场云识别技术研究  

Cloud Recognition in Heliostat field of Tower Concentrating Solar Power Station Based on Feature Algorithm

作　　者：谢昱卓 李刚[1] 倪杭飞 XIE Yuzhuo;LI Gang;NI Hangfei(Mechatronics T&R Institute,Lanzhou Jiaotong University,Lanzhou 730070,China;Zhejiang Supcon Solar Technology Co.,Ltd.,Hangzhou 310053,China)

机构地区：[1]兰州交通大学机电技术研究所,甘肃兰州730070 [2]浙江中控太阳能技术有限公司,浙江杭州310053

出　　处：《智慧电力》2021年第11期38-44,共7页Smart Power

基　　金：国家自然科学基金资助项目(72061021);甘肃省高等学校科研项目资助(2018C-10)。

摘　　要：地表太阳辐照度是塔式光热电站稳定运行的关键影响因素。针对电站镜场云层遮挡会影响地表辐照度数值,带来设备安全与集热量之间的矛盾问题,利用全天空成像仪对镜场上空云层进行实时监测。结合Otsu、Surf和HOG算法,提出颜色、边缘和纹理组合视觉特征算法。根据真实采集云层图像对所提组合算法与单一算法进行准确率对比分析。结果表明,组合算法对各类云状识别准确率远高于单一算法,验证了所提方法的有效性和优越性。Surface solar irradiance is a key factor for influencing the stable operation of tower concentrating solar power(CSP) stations.In view of the fact that cloud occlusion in the heliostat field of the power station will affect the irradiance value of the ground and bring about the contradiction between equipment safety and heat collection,total sky imager is used to get cloud fraction over the heliostat field in real time. The paper proposes a combined visual features algorithm of color,edge,and texture by combining Otsu,Surf,and HOG. After image preprocessing,the accuracy of the combined algorithm and the single algorithm are compared and analyzed according to the real collected cloud image. The results show that the accuracy of the combined algorithm for various types of cloud recognition is much higher than that of the single algorithm,and the effectiveness and superiority of this method is verified.

关 键 词：全天空成像仪 云状识别 OTSU SURF HOG 

分 类 号：TM615[电气工程—电力系统及自动化]