Crop classification using PerceptiveSentinel

Efficient and accurate classification of land cover and land usage can be utilized in many different ways: ranging from natural resource management, agriculture support to legal and economic processes support. In this article, we present an implementation of land cover classification using the PerceptiveSentinel platform. Apart from using base 13 bands, only minor feature engineering was performed and different classification methods were explored. We report an F1 and accuracy score (80-90%) in range of state of the art when using pixel-wise classification and even comparable to time series based land cover classification.