Abdullah ERZIN

Redesigned robotic camera-to-arm calibration, reducing calibration time by over 80% and achieving 5 mm positioning error. Developed fruit-tracking and cluster-aware picking algorithms for autonomous harvesting robots, and led vision integration and field validation across five countries.

Led end-to-end software development of an autonomous solar panel cleaning and inspection robot from concept to functional prototype. Engineered a custom NumPy-based SLAM system and developed A*-based navigation with real-time perception on NVIDIA Jetson Orin.

Engineered a real-time face recognition system on NVIDIA Jetson Nano with 98% accuracy and built a stereo-camera liveness detection module. Implemented performance-critical components in C++ and optimized a YOLO-based video anonymization pipeline.

Developed and deployed industrial machine vision applications using HALCON, C#, and MSSQL, integrating automated inspection workflows into desktop software. Led a TÜBİTAK-funded thermal non-destructive testing research project.

Developed proof-of-concept machine vision applications using HALCON, C#, and SQL, including an OpenCV-based implementation. Participated in industrial site visits to gain practical exposure to camera, lens, and lighting configurations.

Completed a Bachelor of Science in Mechatronics Engineering.