A DATASET AT A GLANCE...

Weather Image Recognition

Be it a simple outdoor activity like, playing an outdoor game or more complex operations that might include: (i) take-off, landing and air traffic management in aerodrome operations, (ii) planning irrigation facilities and seeding of crops in an agricultural environment, (iii) the selling of a utility based on the how hot or cold the climatic condition exits, (iv) the alarming of the public about the probability of occurrence of a forest fire or a heavy rain,  (v) the safe boarding of a ship, (vi) the estimation of the amount of renewable power that can be generated  and many more depend commonly on one thing- “Weather Prediction”. Weather has a great impact on all the happenings that take place in varied regions around the world. Forecasting the weather from real-time signals or images, prior to initiating a weather- related application, can cut down huge losses to life and money.  Researchers, who desire to serve in varied fields like, aerospace. marine, forestry, renewable energy optimization and agriculture can hence use the “Weather phenomenon database (WEAPD)” from Harvard Dataverse (https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/M8JQCR). This dataset can also be downloaded from the Kaggle repository (https://www.kaggle.com/datasets/jehanbhathena/weather-dataset) and it encompasses about 6862 images of varying weather conditions in JPG format to perform a classification task in to  hail, snow, lightning, rainbow, rain, dew, sandstorm , rime, frost, glaze and fog/smog. Though the dataset is publicly available, citation request as “Xiao H , Zhang F , Shen Z , et al. Classification of Weather Phenomenon From Images by Using Deep Convolutional Neural Network. Earth and Space Science, 2021, 8(5). doi: 10.1029/2020EA001604” is expected from the users. To conclude, this dataset is believed to aid in developing systems that render highly accurate weather predictions from the images itself.

Image credit: www. unsplash. com