Geothermal Surface Manifestation photographic dataset

This is a zip file of Geothermal Surface Manifestation (GSM) photographic dataset. It contains 5,000 images, 500 images each type of eight types of GSMs. The eight types of GSMs are warm spring, hot spring, geyser, fumarole, mud pot, hydrothermal alteration, crater lake, and none GSM. These GSM images are used for training and testing a GoogLeNet model to perform recognition of GSMs. The overall accuracy of this model reaches about 90%.

It includes eight folders after unzipping:
1.WS-Warm spring, 
2.HS-Hot spring, 
3.FO-Fountain, 
4.FU-Fumarole, 
5.MP-Mud pot, 
6.HA-Hydrothermal alteration, 
7.CL-Crater lake,
8.NG-No geothermal

It is a benchmark dataset for a manuscript titled "Recognition of surface geothermal manifestations: a comparison of machine learning and deep learning" .