Aggregating remote sensing satellite imagery categorized for small-object detection using an end-to-end edge-enhanced GAN. It features high-resolution image pairs processed through an integrated object detector network to identify minute geographical features.
Use Cases
- Train a super-resolution model using the GAN-enhanced satellite images to improve feature extraction for small objects
- Evaluate object detection performance on low-resolution remote sensing data using the integrated detector network
- Develop edge-enhancement algorithms to sharpen boundaries of small objects in satellite imagery
Strengths
- Features remote sensing satellite imagery specifically for small-object detection
- Includes an End-to-End Edge-Enhanced GAN (EE-GAN) for image preprocessing
- Integrates an Object Detector Network for simultaneous feature enhancement and classification