土石堤坝渗漏红外-可见双光融合图像数据集

针对实际土石堤坝表面凹凸不平、植被覆盖等引起的正常堤坝表面常出现与渗漏异常特征相符的温度分布从而易引起模型及人工基于红外图像出现误判的问题，采用图像融合算法将红外图像同其对应的可见光图像融合使用，生成了既能提供温度信息又能提供真实空间感的融合图像，从而研究对复杂土石堤坝现场环境具有强适应性的基于双光融合图像的堤坝渗漏精细辨识模型。数据量为169M。

Aiming at the issue that uneven surfaces and vegetation cover on actual earth-rockfill dams often result in normal dam surfaces exhibiting temperature distributions consistent with those of abnormal seepage characteristics, which easily leads to misjudgments by both intelligent models and human operators when analyzing infrared images. To address this problem, an image fusion algorithm was utilized to fuse infrared images with their corresponding visible light images, generating fused images that can provide both temperature information and realistic spatial perception. Accordingly, a fine-grained seepage identification model for dams based on dual-light fusion images was developed, which exhibits strong adaptability to the complex on-site environment of earth-rockfill dams. The total size of this dataset is 169 MB.

搜集汇总

背景与挑战

背景概述

该数据集是一个专注于土石堤坝渗漏检测的红外-可见双光融合图像集合，旨在解决传统红外图像因堤坝表面复杂环境（如凹凸不平和植被覆盖）导致的误判问题。通过融合红外和可见光图像，数据集提供了同时包含温度信息和真实空间感的图像，以支持开发强适应性的堤坝渗漏精细辨识模型。数据集规模为166.7MB，包含1069个文件，由河海大学在2023年发布，属于国家重点研发计划项目成果。