Tool to classify cloud into distinct categories, and segment them relative to background (sky)
These instructions will get you a copy of the project up and running on your local machine for development and testing purposes. See deployment for notes on how to deploy the project on a live system.
For Python 3
sudo dnf install python3-numpy
sudo dnf install python3-opencv
pip3 install tensorflow
pip3 install opencv-contrib-python
Import skyweatherCloud library
import skyweatherCloud
Define paths to image, model (Tensorflow Frozen Graph) and label. Example...
file_name = "images/img3.png"
model_file = "model/yaraCloudNet_v1.pb"
label_file = "model/yaraCloudNet_v1.txt"
Create TF Record files from data set
python3 build_image_data.py --train_directory=data/train --output_directory=data \
--validation_directory=data/validation --labels_file=labels.txt \
--train_shards=2 --validation_shards=2 --num_threads=1
Add nets to your Python path
export PYTHONPATH="$PYTHONPATH:/home/ottermegazord/PycharmProjects/yara-cloud/object_detection/models/nets"
Train model
python3 object_detection/legacy/train.py --train_dir='OUTPUT TO YOUR TRAINED MODEL' \
--pipeline_config_path='PATH TO YOUR CONFIG FILE'
Create skyweatherCloud object
cloud = skyweatherCloud.Cloud(file_name, model_file, label_file)
Print cloud classification and segmentation index
print(cloud.pred())
Sample output
['c thick dark', '> 90%']
- Idaly Ali - GitHub
This project is licensed under the MIT License - see the LICENSE.md file for details
- Tensorflow