Land Cover Classification using a Simple Deep Learning Model (TensorFlow) with Earth Engine Data

Hi Geospatial Enthusiast! Would you like to classify land cover classification using a deep learning model such as Convolution Neural Network? In this video, I will show you how to use Earth Engine imagery and samples to build a land cover classification model in TensorFlow (Keras). GEE script: https://code.earthengine.google.com/4... Colab script: https://colab.research.google.com/dri... The image used is the 2023 Landsat 8 and 9 composite of Jambi City, Indonesia. The land cover classes I use are built-up, bare land, water, wetland, herbaceous, dry shrub, wet shrub, palm oil, and plantation forest. Other than the spectral band from Landsat, the features or variables I use are the spectral indices and elevation from SRTM. The images/variables and extracted samples will then be exported to Google Drive where they will be mounted and used in Colab. For the model, I am using a combination of Dense and Convolutional 1-dimensional layers. My model is far from perfect, you could improve it more. Also, this is 1 dimensional (tabular), you could use more dimensions using spatial data instead of extracted tabular data. I hope this video helps with your research or project. Email: ramiqcom@gmail.com Linkedin:   / ramiqcom   GitHub: https://github.com/ramiqcom