刘志清, 兰奇逊, 徐林, 李亚杰, 王朝勇. 基于图像处理的光伏局部阴影区域分割方法[J]. 太阳能学报, 2023, 44(2): 391-398. DOI: 10.19912/j.0254-0096.tynxb.2021-1046
引用本文: 刘志清, 兰奇逊, 徐林, 李亚杰, 王朝勇. 基于图像处理的光伏局部阴影区域分割方法[J]. 太阳能学报, 2023, 44(2): 391-398. DOI: 10.19912/j.0254-0096.tynxb.2021-1046
Liu Zhiqing, Lan Qixun, Xu Lin, Li Yajie, Wang Chaoyong. PARTIAL SHADING REGION SEGMENTATION OF PHOTOVOLTAIC SYSTEMS BASED ON IMAGE PROCESSING[J]. Acta Energiae Solaris Sinica, 2023, 44(2): 391-398. DOI: 10.19912/j.0254-0096.tynxb.2021-1046
Citation: Liu Zhiqing, Lan Qixun, Xu Lin, Li Yajie, Wang Chaoyong. PARTIAL SHADING REGION SEGMENTATION OF PHOTOVOLTAIC SYSTEMS BASED ON IMAGE PROCESSING[J]. Acta Energiae Solaris Sinica, 2023, 44(2): 391-398. DOI: 10.19912/j.0254-0096.tynxb.2021-1046

基于图像处理的光伏局部阴影区域分割方法

PARTIAL SHADING REGION SEGMENTATION OF PHOTOVOLTAIC SYSTEMS BASED ON IMAGE PROCESSING

  • 摘要: 为解决真实场景中光伏组件局部阴影区域分割受环境干扰的问题,该文提出基于图像处理的光伏局部阴影区域分割方法。根据光伏组件特征用改进的阈值分割方法提取组件边框,采用Hough变换检测组件边框线段,计算包围边框线段端点的最小凸包作为感兴趣区域。在感兴趣区域内再次使用改进的阈值分割方法提取局部阴影区域。用模拟仿真图像和真实场景图像验证该文所提方法的有效性,实验结果表明:所提方法可准确分割出光伏组件上的局部阴影区域且保持了遮挡区域的细节信息。与多阈值分割方法和基于Canny边缘检测的分割方法相比,该文所提方法的错误分类误差更小,类别像素准确率更高。

     

    Abstract: In order to solve the problem that the partial shading region segmentation of photovoltaic(PV)modules disturbed by the environment in the real scene,an image processing based partial shading region segmentation of PV systems is proposed in this paper.According to the characteristics of PV modules,firstly the module frame is extracted by the improved threshold segmentation method,secondly the module frame lines are detected by Hough transform,then the minimum convex hull surrounding the endpoints of the frame lines is calculated as the ROI. The partial shading region in the ROI is extracted by the improved threshold segmentation method. The effectiveness of the proposed method is verified by simulated images and real scene images. The experimental results show that the proposed method can accurately segment the partial shading region on the PV module and maintain the details of the shading region.Compared with multiple threshold segmentation method and Canny edge detection based segmentation method,this method has smaller misclassification error and higher class pixel accuracy.

     

/

返回文章
返回