Cyclone Detection

Goal 🎯 : The main objective of this project is to detect the intensity of cyclones happening all around the world using satellite image data.

DATASET ℹ️ : https://www.kaggle.com/datasets/sshubam/insat3d-infrared-raw-cyclone-images-20132021

DESCRIPTION: The project consists of an image dataset of cyclones 🌀 containing infrared and raw satellite images from INSAT-3D. It also consists of a .CSV file containing the intensity of the cyclones. It comprises a model built using the Inception_v3 model.

WHAT I HAD DONE :

1. Firstly, I had seen the composition of the dataset and came to know it consists of a .csv file and IR and raw satellite images.
2. Imported the necessary libraries and modules to perform data visualization and build the model.
3. Plotted the 25 Images from the Infrared and Raw Images Dataset using matplotlib.pyplot..Also, the conclusion image is derived from each image (present in the dataset).
4. Build a function to load the images using the load_img. function and store it in a list in form of np.ndarray..
5. I created the training dataset using the ImageDataGenerator. and flow_from_directory. functions.
6. Model Building: After creating the training dataset, I build the model using Inception_v3 (Pre Trained model) from tensorflow.keras.applications.inception_v3.. Then, I evaluated my model by .evaluate(). and got a loss of 1.3165.
7. Finally, I saved the model using save_model. from the Keras models. module.

LIBRARIES NEEDED:

1. Numpy

2. Pandas

3. Matplotlib

4. Tensorflow

CONCLUSION: After observing the model, it can be inferred that the model incurred loss (Mean Absolute Error) of around 1.3165 on the training data which is considerably good. At last, model is successfully detecting the intensity of a cyclone from images.

Contributed By : **ANKIT PANDEY You can Connect with me at : LinkedIn : www.linkedin.com/in/ankit-pandey-2003ap Twitter : https://twitter.com/thecoolpandey