基于LeNet-5模型太阳能电池板缺陷识别分类.PDFVIP

第XX 卷 第XX 期 热 力 发 电 Vol.XX No.XX 选择年份 选择月份 THERMAL POWER GENERATION 选择月份 选择年份 基于LeNet-5 模型的 太阳能电池板缺陷识别分类 吴 涛，赖 菲 （西安热工研究院有限公司，陕西 西安 710054） ［摘 要］太阳能电池板是光伏发电组件的核心部件，其质量的优劣直接关系安全发电和发电效率。 因此，对太阳能电池板进行缺陷检测具有重要的实际价值。考虑到人工检测的低效性和高 成本，本文提出利用在深度学习领域图像分类性能良好的卷积神经网络对太阳能电池板图 像进行自动识别分类。利用Tensorflow 平台Tensorboard 的可视化性能，对经典LeNet-5 模 型进行结构改善和超参数的调整，并将改进LeNet-5 模型与经典LeNet-5 模型和支持向量 机的分类结果互相对比，结果表明改进LeNet-5 模型的分类效果最优。 ［关 键 词］太阳能电池板；LeNet-5 模型；图像分类；神经网络；超参数；Tensorboard ［中图分类号］TP391.4 ［文献标识码］A ［DOI 编号］10.19666/j.rlfd.201807147 ［引用本文格式］吴涛, 赖菲. 基于LeNet-5 模型的太阳能电池板缺陷识别分类[J]. 热力发电, 年, 卷(期): 起始页码-终止页 码. WU Tao, LAI Fei. Identification and classification of defects in solar cells based on LeNet-5[J]. Thermal Power Generation, 年, 卷(期): 起始页码-终止页码. Identification and classification of defects in solar cells based on LeNet-5 WU Tao, LAI Fei (Xi’an Thermal Power Research Institute Co., Ltd., Xi’an 710054, China) Abstract: Solar cell, as the core component of photovoltaic power generation module, its quality directly related to security and efficiency of power generation. So the defect detection of solar cells plays a very important significant role. Taking into account the inefficiencies and high costs of manual , this paper proposes to use the convolutional neural network with good image classification performance in the deep learning field to automatically identify and classify the cell. In order to adapt LeNet-5 to this classification, this paper based on the visual performance of tensorboard of tensorflow platform to improve its structure and hyper parameters, and compare it with the classic LeNet-5 an