Kursat

Komurcu

Currently working on "satellite image change detection" project and publishing papers about it. 1. Satellite image change detection using segmentation and VAR model 2. Zero Shot Classification for change detection in satellite imagery (OPENAI CLIP Model) 3. Caption generation for satellite images and classify changes according to captions (Llama and traditional ML algorithms) 4. Measuing performance of Llama model for satellite image object recognition and compared it with other results.

1. Semi-Supervised traffic object detection with drone data 2. Landmine detection with drone 3. Object tracking with drone using ROS

Secured 2nd place in algorithm development in an autonomous vehicle competition which organized by Turkish Defense Industry Presidentship. Advanced slope estimation and obstacle mapping with Lidar, enhancing navigational safety. Detected moving obstacles using 3D camera technology. Created a user-friendly QT-based GUI for efficient autonomous robot management. Developed an HD Map node for superior navigation precision and spatial awareness. Implemented a precise Lane Detection node, ensuring accurate vehicle lane adherence.