Satellite Imagery Feature Detection using Deep Convolutional Neural Network: A Kaggle Competition

This is a 2017 competition entry for a DSTL Kaggle challenge. While it represents a solid technical approach for its time, it has zero current community traction (0 stars) and the techniques described (FCN-based segmentation) have since been superseded by more advanced architectures like U-Net variants, SegNet, and Vision Transformers. Frontier labs and specialized geospatial companies now offer these capabilities as commodity features.