Back to projects
Model DevelopmentFeature ExtractionData Engineering

By & Havn Sea Traffic Monitoring

A harbor-monitoring computer vision system for detecting vessel type and direction from real Copenhagen camera streams.

Role
Computer vision and deployment engineering
Focus
Model Development / Feature Extraction
Stack
Python, FastAPI, Docker

Context

The shape of the problem

The project sat at the intersection of urban operations and messy real-world video. Harbor cameras see weather, occlusion, reflections, and many vessel types, so the task was less about a pristine model demo and more about making detection useful under practical deployment constraints.

Build

What I actually built

I worked on the model pipeline, training flow, and deployment structure around PyTorch, AzureML, Kubernetes, and CI/CD. The system had to classify vessel types and sailing direction while remaining maintainable enough for future retraining and broader rollout.

Engineering

Key decisions

01

Ensuring robust vessel detection under varying weather and lighting conditions

Refactored legacy codebase and modularized key components for maintainability

02

Accurately classifying diverse ship types in busy maritime environments

Trained custom computer vision models using PyTorch to detect vessel types and motion

03

Tracking vessel direction and motion with high reliability

Integrated with AzureML for cloud-based training and experimentation

04

Improving the maintainability and scalability of the legacy system

Streamlined Kubernetes deployment with improved CI/CD setup for reliable retraining and rollout

Outcome

What changed

  1. 01Achieved over 90% accuracy in detecting correct vessel types and sailing directions
  2. 02Improved system robustness and maintainability with modular microservice structure
  3. 03Enabled scalable cloud training via AzureML for future iterations
  4. 04Prepared foundation for broader implementation across the city's harbor surveillance infrastructure

Stack

Tools and systems

PythonFastAPIDockerAzureMLAzure DevOpsAzure Kubernetes Service

I build delightful systems.

© 2026 Kristian Ernst.