Understanding land use from satellite images
The 16-session plan
Research existing AI models that classify land use from satellite images and summarize key findings in a presentation. Analyze the EuroSat dataset, identifying the categories of land use represented and creating a data organization plan.
Begin developing a convolutional neural network (CNN) in Python using TensorFlow to classify the satellite images based on the identified categories. Implement data augmentation techniques on the dataset to improve the robustness of the AI model.
Implement a convolutional neural network (CNN) using TensorFlow to begin training on the EuroSat dataset for land use classification. Fine-tune the CNN by experimenting with different hyperparameters and record the accuracy of the model on a validation set.
Frequently asked questions
Who is this project a great fit for?+
BlueRobins projects are designed for curious students in Grades 3-12 who want to build. No prior experience is needed — mentors start wherever your child is and grow the challenge as they grow.
What does a typical session look like?+
Your child meets 1-on-1 with their mentor every week over Google Meet. Programs run 16 sessions of 45 minutes each. Sessions are hands-on from minute one — building, designing, coding, writing, experimenting — with the mentor coaching live, asking great questions, and unlocking the next step.
What will my child build by the end?+
A finished project they're proud of and can show the world. Across 16 sessions, your child turns an ambitious idea into a working app, prototype, research write-up, portfolio, or published artifact. Every student finishes with a completion certificate signed by their mentor.
Why a free trial first?+
So your child can meet their mentor, get a feel for the project, and leave the call excited. The trial is a complimentary 1-on-1 free trial session — no credit card needed. After the trial, you decide if it's the right fit. Most families know within the first 15 minutes.