基于改进CNN的太阳能板遮挡物分类研究

Application of improved CNN in the classification of solar panel shelters

我国西北地区的太阳能发电站分布广泛,太阳能板的维修和维护耗时耗力.本文采用基于改进CNN算法,搭建模型进行太阳能板遮挡物分类研究.首先对太阳能板图像数据进行采集和预处理,利用改进CNN算法对太阳能板图像进行训练,得到训练模型,然后对太阳板遮挡物图像进行分类,算法总体的分类准确率高达97.65%.实验表明,本算法在太阳能板遮挡物分类方面有较好的实用性.

With the wide distribution of solar power stations in northwest China,the repair and maintenance of solar panels are time-consuming and labor-intensive.This paper uses the improved CNN algorithm to build a model to study the classification of solar panel occlusion.First,collect and preprocess the solar panel image data,train the solar panel image using the improved CNN algorithm to obtain the training model,and then classify the solar panel occlusion image.The overall classification accuracy of the algorithm is as high as 97.65%.Finally,the experiment shows that,This algorithm has good practicability and research value in solar panel occlusion classification.