ELPV

该数据集是在光伏电池异常检测领域中被广泛使用的一份文献资料，特别适用于利用电致发光图像对光伏电池缺陷进行标注和检测。该数据集包含经四类分类标注的图像，这些分类依据缺陷可能性进行划分，并为训练目的将其图像大小调整为640×480像素。规模上，该数据集包含2352张经过增强处理的单晶光伏电池图像，原始图像数量为588张。其任务是对光伏电池图像中的异常检测和分割。

This dataset is a widely utilized literature resource in the field of photovoltaic cell anomaly detection, and is particularly applicable for labeling and detecting defects in photovoltaic cells via electroluminescence images. The dataset comprises four-category annotated images, which are classified based on defect probability, and all images have been resized to 640×480 pixels for training purposes. In terms of scale, this dataset contains 2352 augmented monocrystalline photovoltaic cell images, originally derived from 588 raw images. Its core task is anomaly detection and segmentation for photovoltaic cell images.

搜集汇总

背景与挑战

背景概述

ELPV数据集是一个包含2,624个太阳能电池图像的集合，用于视觉识别缺陷太阳能电池的研究。数据集提供了每个图像的缺陷概率和电池类型标注，适用于机器学习和计算机视觉应用。