african-ai/challenge2_ViewFromTheTop

# View From The Top : Feature Detection of Aerial Imagery ## Data Challenge Task ###### classifying a high resolution aerial image into 9 types of urban land cover; trees, grass, soil, concrete, asphalt, buildings, cars, pools, shadows ## Data Challenge Bounty ### AWS Credits Worth USD 2500 to the **two** highest ranked model submissions, based on Git timestamps credits valid till February 2019, courtesy of [Amazon Web Services] ### USD 100 to the **third** highest ranked model submission, based on Git timestamps Prize Money Courtesy of [Africa's Talking] ## Due Date: 7th September 2018 ## Why is this Important? Maps are absolutely essential for decision support. Knowing where buildings are located is a fundamental input for urban planning, public safety, public health, disaster response, environmental protection, sustainable development and census data, among other examples. Some of these applications typically require timely and high-resolution maps. Increasingly, aerial imagery of larger areas can be covered using small commercial mapping drones. As such, aerial imagery is quickly becoming a Big Data problem. A single 20-minute drone flight can capture 800 high-resolution images, for example. Academic research has shown that trained analysts will take ~ 1 minute to analyze a single high-resolution image. Manually analyzing 800 images from a 20-minute flight can therefore take 13 hours. ###### Rules of Engagement The Data Challenge is judged based on the following criteria: - A Correct fork, branch and pull request - Using the GitHub Pull Request timestamp where order of submissions is applicable - Using solution quality/accuracy and explanation to rank submissions where applicable - Do not share any code that you cannot open source on the Git Repository as its open source and african.ai will not be liable for any breach of intellectual property (if at all) once shared on the platform. ## Working on the Data Challenge 1.Fork the code challenge repository …