Thermal Image dataset for object classification

The thermal image dataset for object classification is comprised of total 6414 images were captured by using Seek Thermal and a total of 1014 images are captured by using FLIR. The resolution of the images captured by the FLIR camera is 1080 X 1440 and for Seek Thermal is 300 x 400. There are three classes cat, car and man. The resolution of the images is higher in comparison with the handheld thermal cameras. The texture of the object is not clear but the temperature of the object is distinguishable, edges and contours are also clear. A thermal image dataset for object classification is used for the training and testing of deep learning models.

本目标分类热成像数据集（thermal image dataset for object classification）共包含6414张由Seek Thermal拍摄的图像，以及1014张由FLIR拍摄的图像。FLIR相机拍摄的图像分辨率为1080×1440，Seek Thermal拍摄的图像分辨率为300×400。数据集共涵盖猫、汽车、行人三类目标。该数据集的图像分辨率高于普通手持热成像相机的输出分辨率。图像中物体的纹理虽不够清晰，但物体的温度特征可被有效区分，边缘与轮廓则较为明晰。本数据集可用于深度学习模型的训练与测试。

搜集汇总

背景与挑战

背景概述

该数据集包含7428张热成像图像（Seek Thermal设备6414张，FLIR设备1014张），分辨率分别为300x400和1080x1440，涵盖猫、汽车和人三个类别，适用于深度学习模型的训练和测试。