japhethkimeu / deforestation_detection_using_deep_learning

Project on detecting deforestation using deep learning algorithms

Detecting deforestation using deep learning

This projet entails the application of deep learning algorithms i.e. CNN to detect deforestation using satellite imagery.

Data

I analysed Sentinel-2 images (cloud free) of three different regions in the State of Mato Grosso in Brazil and derived a total of 5,122 image chips and masks of size 256 * 256. See a sample of training images and masks in the sample data folder.

Data Augmentation

Image chips were flipped left and right before fed into the network for training. See preprocess.py

Model

Unet model was implemented by using MobileNetV2 as the feature extraction with pre-trained weights.

More on Unet model U-Net: Convolutional Networks for Biomedical Image Segmentation

Training

The model was trained first using the pre-trained weights of the base model for 15 epochs with early stopping. The best result was achieved by re-training the whole model for 20 epochs and attained a binary accuracy of 0.9780.

Requirements

    - tensorflow
    - keras
    - numpy
    - matplotlib
    - pandas
    - cv2

Results

deforestation_detection_using_deep_learning's Issues

send link of original dataset

sample dataset is not giving accurate results ,please do send the original dataset

Steps on Generating and Saving Training Data for Deforestation Deep Learning Project

I'm facing challenges in successfully running the project due to the data setup. The data you've used is gathered from Google Drive, with images and masks stored in separate folders ("images" and "masks"). I'm particularly interested in generating the data from Sentinel 2 image collection and exporting this data to the drive to be compatible with your codebase, but I'm uncertain about the exact steps to follow. I really appreciate your time and attention to this issue. Thanks.